Disrupted development and imbalanced function in the global neuronal workspace: a positive-feedback mechanism for the emergence of ASD in early infancy

Chris Fields; James F Glazebrook

doi:10.1007/s11571-016-9419-8

. 2016 Nov 15;11(1):1–21. doi: 10.1007/s11571-016-9419-8

Disrupted development and imbalanced function in the global neuronal workspace: a positive-feedback mechanism for the emergence of ASD in early infancy

Chris Fields ^1,^✉, James F Glazebrook ²

PMCID: PMC5264757 PMID: 28174609

Abstract

Autism spectrum disorder (ASD) is increasingly being conceptualized as a spectrum disorder of connectome development. We review evidence suggesting that ASD is characterized by a positive feedback loop that amplifies small functional variations in early-developing sensory-processing pathways into structural and functional imbalances in the global neuronal workspace. Using vision as an example, we discuss how early functional variants in visual processing may be feedback-amplified to produce variant object categories and disrupted top-down expectations, atypically large expectation-to-perception mismatches, problems re-identifying individual people and objects, socially inappropriate, generally aversive emotional responses and disrupted sensory-motor coordination. Viewing ASD in terms of feedback amplification of small functional variants allows a number of recent models of ASD to be integrated with neuroanatomical, neurofunctional and genetic data.

Keywords: Categorization, Connectome, Predictive coding, Prenatal development, Resting-state networks, Small-world networks

Introduction

Autism spectrum disorders (ASD) are early-onset conditions characterized by deficits or delays in social interaction and language, repetitive, stereotypical behaviors and restricted interests (DSM-5; American Psychiatric Association 2013). They have high (>1%) prevalence with an approximately 4:1 bias toward males, varying severity, and unknown, apparently heterogeneous etiology (reviewed by Volkmar et al. 2005; Geschwind 2009; Leonard et al. 2010; Matson and Kozlowski 2011) and are often comorbid with other psychiatric conditions ranging from Fragile X syndrome to epilepsy (reviewed by Jeste 2011; Peca and Feng 2012). Dozens of genes representing multiple functional families have been associated with ASD (reviewed by Geschwind 2008; Betancur 2011; Miles 2011; Geschwind and Flint 2015). Heritability is high (40–80%; Geschwind and Flint 2015, Fig. 1), but de novo mutations including de novo copy-number variants are also common. Functional-network analysis confirms that the products of many ASD-associated genes are involved in neuronal growth and synaptic regulation (Chow et al. 2012; Parikshak et al. 2013; Pinto et al. 2014); in some cases, ASD-relevant phenotypes have been directly confirmed in animal models (e.g. Vaccarino et al. 2009). Gene expression studies indicate that at least some ASD-related gene products are active in early- to mid-fetal development (Parikshak et al. 2013; Willsey et al. 2013). A wide variety of dietary and other environmental factors expected to affect nervous-system development have also been associated, either prenatally or postnatally, with ASD (reviewed by Currenti 2010; Gardener et al. 2011). Functional connectivity studies in both children and adults with ASD have reported both under-connectivity, particularly between hemispheres, and over-connectivity, especially locally; however, the blanket statement that ASD is characterized in general by local overconnectivity and global underconnectivity is increasingly not supported (reviewed by Courchesne and Pierce 2005; Geschwind and Levitt 2007; Minshew and Williams 2007; Rippon et al. 2007; Müller et al. 2011; Wass 2011; Vissers et al. 2012; Maximo et al. 2014; Tyszka et al. 2014; Picci et al. 2016). A recent study of high-functioning ASD adults, for example, demonstrated both significant differences in the spatial patterns of inter-hemispheric functional connections in ASD compared to typical subjects and significantly greater functional connection heterogeneity among the ASD group, leading to the suggestion that functional connectivity patterns in ASD are idiosyncratic, not stereotypical (Hahamy et al. 2015).

Fig. 1 — a A simple SWN with high modularity, i.e. clearly distinct “communities” of nodes. “Hubs” are nodes through which connections between modules pass. Note that each module contains two *triangles*. b A modular network with no triangles and hence a clustering coefficient of zero; this network is not an SWN. c A less-modular SWN with more *triangles* and hence a higher clustering coefficient than the network in a

Early models of ASD tended to focus on one or more of the primary symptomatic presentations (for an extensive historical review, see Rajendran and Mitchell 2007). Examples include models organized around deficits in social cognition (Baron-Cohen 2002), aversion to novelty (Markram et al. 2007) or weak central coherence, i.e. a relative inability to grasp overall gestalt features of a scene or situation (Happé and Frith 2006). The explosion of data on the functional neuroscience of ASD since the early 2000s has motivated a second generation of models that focus not on overt symptoms but on proposed developmental and neurocognitive mechanisms. An early example is the proposal of Williams et al. (2001), later elaborated by Iacoboni and Dapretto (2006) and Oberman and Ramachandran (2007), that the social-cognition deficits typical of ASD are the result of mirror-system dysfunctions. Along similar lines, Pelphrey et al. (2011) proposed that early dysfunction in posterior superior temporal sulcus (STS), a component of the mirror system, induced dysfunction throughout the “social brain” network and hence dysfunctions in social cognition. More recently, a number of authors have proposed second-generation models of ASD as primarily disorders of perceptual categorization and/or the generation and evaluation of perceptual expectations, and have shown that the social interaction, language, and other typical presentations of ASD can be explained on this basis (Fields 2012a; Pellicano and Burr 2012; Lawson et al. 2014; Van de Cruys et al. 2014; Hellendoorn et al. 2015). These models suggest a number of different underlying mechanisms: Fields (2012a) attributes ASD to disrupted category learning, Pellicano and Burr (2012) to attenuated top-down perceptual expectations, Lawson et al. (2014) and Van de Cruys et al. (2014) to overly-precise representations of differences between expectations and perceptions, and Hellendoorn et al. (2015) to deficits in the detection of perceptual invariances.

While existing second-generation models have viewed ASD in primarily cognitive or cognitive-affective terms, both developmental and clinical perspectives suggest an expanded and more fully integrated view of ASD that explicitly recognizes the role of variant experiences of the world, beginning in earliest infancy, in generating variant behavioral responses that, in turn, provide variant contexts for further experiences. Such an expanded view suggests a focus not on particular processes or modules but on the coordination of neural activity across multiple functional networks. Overall functional coordination of both experience and behavior is generally attributed to the global neuronal workspace, henceforth abbreviated to GNW (Dehaene and Naccache 2001; Dehaene and Changeux 2004, 2011; Dehaene et al. 2014), also referred to as the global workspace (Baars 1998; Baars and Franklin 2003; Baars 2005; Baars et al. 2013), connective core (Shanahan 2012) or “rich club” (Sporns 2013), a network of long-distance cortico-cortical and cortico-thalamic pathways. Normal GNW function depends critically on how the total connectivity of the network is allocated between typically shorter-range connections within networks subserving particular functions and typically longer-range connections between such networks. As suggested by Belmonte and Baron-Cohen (2004), disruptions of this balance between within- and between-network connectivity can be expected to present symptomatically as cognitive and/or emotional disorders including ASD (Just et al. 2012; Peters et al. 2013; Glazebrook and Wallace 2015). New data on the functional organization and connectivity of the GNW prenatally through the first two years (reviewed by Di Martino et al. 2014; Dehaene-Lambertz and Spelke 2015; Vértes and Bullmore 2015) provide a basis for examining a potential role for such GNW imbalances in the critical early-developmental stage of ASD.

Motivated by these new data and theoretical approaches, we here review theoretical, computational and experimental studies which together suggest that ASD is characterized, beginning in earliest infancy, by a positive feedback loop that amplifies small functional variations in early-developing sensory processing and sensory-motor pathways into structural and functional imbalances in the developing GNW. This suggestion is consistent with the functional development of the nervous system being fundamentally interactive (reviewed by Johnson 2011) and with the emerging view of developmental neuropsychiatric disorders in general as disruptions of typical “connectome” development (Di Martino et al. 2014; Dehaene-Lambertz and Spelke 2015; Vértes and Bullmore 2015). Using the well-established graph-theoretic model of the GNW as a small-world network, henceforth abbreviated to SWN (Sporns et al. 2002; Sporns and Zwi 2004; Achard et al. 2006; Bassett and Bullmore 2006; Sporns and Honey 2006; Glazebrook and Wallace 2009; reviewed by Rubinov and Sporns 2010; Sporns 2013), we show how small variations in the relative probabilities of forming stable within- and between-network connections can have dramatic outcomes for overall GNW function. The broad perspective obtained by considering the effects of local functional perturbations on the GNW as a whole allow us to integrate the mechanistic and etiological proposals of multiple, prima facie competing models within a single developmental framework. Within this feedback-amplification framework, both structural and functional heterogeneity of outcomes is to be expected; we consider this to be a major contribution of the proposed approach.

While it is increasingly clear that multiple sensory systems as well as multi-sensory integration are affected in ASD (reviewed by Baum et al. 2015), both typical and variant visual processing, including the cognitive and affective interpretation of visual input, are the most well-characterized and recent models mainly reference these processes. We therefore focus on early-developmental interactions between visuomotor, medial visual/categorization, dorsal and ventral attention, and affective/reward networks, which together implement both the cognitive and affective interpretation of visual input and the use of visual input to control object-directed behavior, and the interaction of these networks with later-developing (Casey et al. 2005) executive and default (i.e. self and self-other representation, see Andrews-Hanna et al. 2014 for review) systems. We examine the expected effects of imbalancing the coordination of these networks by the relative over- or under-production or enhanced or reduced strength of stable within- or between-network connections. We describe how multiple distinct GNW imbalances can be expected to result, given exposure to a typical infant environment, in cognitive-affective outcomes typical of ASD. We then show that the functional disruptions proposed to be associated with ASD by Fields (2012a), Pellicano and Burr (2012), Lawson et al. (2014), Van de Cruys et al. (2014) and Hellendoorn et al. (2015) can be characterized within this integrative framework. Finally, we comment on further extensions of the model and the possibility of direct experimental tests.

“Background: the GNW as an SWN of SWNs” briefly reviews the common structure of small-world networks, evidence that the GNW is an SWN, and caveats that must be borne in mind when describing the GNW as an SWN. “Development of the GNW” briefly reviews the cognitive, affective and behavioral phenomenology with which models of GNW development must be consistent, and considers how new data, primarily on the functional integration between resting-state networks (RSNs) from the prenatal period through late infancy, further constrain such models. A formal framework is then developed that allows the representation of both typical and variant GNW development. “Perturbations of GNW development” distinguishes global from localized disruptions of GNW function, focusing on those that can be expected to be amplified in the typical infant environment. “Five recent models of ASD from a GNW perspective” discusses the functional disruptions advanced by specific recent models within the GNW framework. The paper concludes with the hypothesis that multiple localized and possibly relatively minor perturbations of GNW function could be amplified, in the typical infant environment, to symptomatic ASD.

Background: the GNW as an SWN of SWNs

The mathematical concepts and tools of complex network analysis are increasingly being applied to anatomical, comparative and functional connectome data (Rubinov and Sporns 2010; Sporns 2013). A key concept of such analysis is the SWN (Watts and Strogatz 1998; see also Barabasi and Albert 1999; Newman 2003), defined as a network with both a small average distance between nodes (formally, an average distance logarithmic in the number of nodes) and a high clustering coefficient (formally, higher than the clustering coefficient expected for a random graph). Informally, SWNs are characterized by clusters of locally-connected nodes (also called “communities” e.g. in Sporns 2013) surrounding a relatively small number of mutually-connected hubs, as shown in Fig. 1. However, not all networks containing hubs are SWNs; networks containing no closed triangles, in particular, have clustering coefficients of zero and are therefore not SWNs (Fig. 1b). Reducing modularity by adding redundant paths between clusters render an SWN more robust, i.e. less prone to disconnection by the removal of a single node or connection (Fig. 1c).

Networks with the characteristics of SWNs can clearly be sought and found at many different scales in nervous systems (Sporns and Honey 2006). Examining connectivity across the entire brain either anatomically by tracing white-matter tracts or functionally by measuring correlated activity yields an SWN in which the nodes are anatomical regions with relatively well-defined, coherent functions (e.g. primary visual cortex, medial-temporal area or thalamus) and the connections link these areas into large-scale functional networks (e.g. dorsal and ventral attention systems, default and task networks, memory and affective/reward systems; reviewed by Park and Friston 2013; Rugg and Vilberg 2013; Sporns 2013; Andrews-Hanna et al. 2014; Rolls 2015; Vossel et al. 2014). Taken together, these large-scale functional networks constitute the GNW (Shanahan 2012; Baars et al. 2013; Dehaene et al. 2014), with the most highly mutually-connected hubs within the GNW forming the “rich club” or connective core of the human connectome (van den Heuvel and Sporns 2011; Shanahan 2012; Sporns 2013). The GNW concept was originally introduced to explain both the apparent unity of and the requirement for serial processing in ordinary, waking consciousness (Baars 1998; Dehaene and Naccache 2001; Baars and Franklin 2003; Dehaene and Changeux 2004; Wallace 2005). Functional studies have now largely confirmed the correlation between coordinated patterns of GNW activity and waking consciousness (Baars et al. 2013; Dehaene et al. 2014), although the implementation of conscious awareness and consciously-directed activity is by no means settled (e.g. Block et al. 2014; Graziano 2014; Hoffman and Prakash 2014; Tononi and Koch 2015; van Leeuwen 2015).

While it is standard to conceptualize the GNW as a network of functionally-coherent anatomical regions connected by white-matter tracts (e.g. Park and Friston 2013, Fig. 2b; Sporns 2013, Fig. 3b), doing so can obscure two important facts. First, each functional region is itself a complex network and possibly an SWN. Second and more importantly, inter-regional fiber tracts comprise many parallel connections between individual neurons, many if not most of which implement unique functional outcomes at the whole-network scale. Such complexity is evident at the voxel scale from which regional networks are constructed experimentally (e.g. Park and Friston 2013, Fig. 2d; Sporns 2013, Fig. 3a). Network “hubs” do not, therefore, exchange functionally-uniform signals and the “connection strength” between hubs cannot be regarded as a measure of the strength of any given signal transmitted between functional regions. The situation may instead be as illustrated in Fig. 2, with as least as many functionally-distinct connections between hubs as there are nodes within a hub.

Fig. 2 — “Hubs” in the GNW are themselves complex networks that may be connected by many parallel, functionally-distinct pathways. The “strength” of connection between hubs is not, therefore, a measure of the strength of any given functional connection

Fig. 3 — Schematic representation of GNW development to from mid-fetal development to birth. a Genetically or physically demarcated regions (*dashed circles*) develop within-region connections by late 2nd trimester. b Multi-regional functional networks are sufficiently developed by early 3rd trimester to support pre-term cognitive abilities, but between-network connections are under-developed. c By normal-term birth, within-network connections are robust and “rich club” hubs through which networks are mainly linked have formed (cf. Gao et al. 2015)

As an example, consider the relation between visual object categorization and motor plans for object manipulation, a relationship that has been extensively studied in the case of tool identification and tool use (reviewed by Vingerhoets 2014; Gallivan and Culham 2015; see also Brandi et al. 2014). In typical adults, the medial visual system is able to recognize dozens to hundreds of tools suitable for a broad array of applications, while the frontal-parietal action network is able to plan actions using these tools in multiple orientations to accomplish many different goals (reviewed by Lewis 2006; Brandi et al. 2014; Ishibashi et al. 2016). Tools, tool-related actions and tool-use goals are each elements of complex, at least quasi-hierarchical categories with many semantic and logical relationships between elements. Hammers and screwdrivers, for example, generally serve different functions, and may have similar or quite different sizes, shapes, weights, and locations in the toolbox. A large, heavy screwdriver may substitute functionally for a hammer in some applications, while a hammer being useful as a screwdriver is unlikely. Most people are far more familiar and competent with some tools than others, indicating that connection strengths are highly variable across representations. The robust human ability to improvise tools, invent new tools and adapt existing tools to novel applications shows, moreover, that the entire tool-use network is highly adaptable. The connections between medial-visual representations of tools as meaningful objects and frontal-parietal representations of tool-use goals and actions appear, therefore, to implement many-to-many mappings between complex, at least quasi-hierarchical, and highly-plastic representational domains. How these many-to-many mappings function in contexts involving multiple tools and multiple goals, whether they are better conceptualized in terms of memory or problem-solving and how they degrade in tool-use apraxias remain subject to considerable debate (e.g. Osiurak et al. 2013; Sunderland et al. 2013; Buxbaum et al. 2014).

Three other issues complicate any simple representation of large-scale networks in general and the GNW in particular as SWNs. First, both within-region and inter-regional connections may be either excitatory or inhibitory. Second, the information transmitted along either excitatory or inhibitory inter-regional connections may be complicated functions of local activity. Information-theoretic considerations, for example, suggest that nervous systems capable of flexible behavior in changing environments would ubiquitously employ predictive coding schemes in which model-based expectations are transmitted top-down and prediction errors, i.e. perception–expectation differences, are transmitted bottom-up (reviewed by Friston 2010). Both local and inter-regional functional network analysis provide evidence that mammalian nervous systems are in fact organized in this way (reviewed by Gómez and Flores 2011; Bastos et al. 2012; Shipp et al. 2013; Adams et al. 2015). Third, local integration of either excitatory or inhibitory signals requires signal synchrony or, at a finer scale, temporal coding. Activity in the GNW, in particular, involves global coordination in time of activity across all contributing components of the network (Baars et al. 2013). Analytic measures that consider only a single path between network components, such as the widely-employed “efficiency” measure of Latora and Marchiori (2001), cannot capture this temporal coordination requirement and may therefore yield spurious or misleading results. However, including connection strengths as well as path lengths (Rubinov and Sporns 2010) in the analysis may compensate, if only implicitly, for this weakness, as suggested by both the heritability of some network efficiency variants reported by Fornito et al. (2011) and the correlations between connection efficiency and ASD severity reported by Lewis et al. (2014; see also below).

With these caveats and complications in mind, we follow Shanahan (2012) in conceptualizing the GNW as an SWN in which all or at least most of the contributing “communities” are themselves SWNs, the connections between communities are many-to-many as shown in Fig. 2, predictive coding is employed at multiple scales, and signal synchrony is critical to coherent functioning. McCall and Franklin (2013) have shown this conceptual approach is computationally feasible within the Learning Intelligent Distribution Agent (LIDA) architecture, a hybrid computing architecture that models the GNW (Franklin et al. 2007); Grossberg and Seidman (2006) have obtained similar results using the adaptive resonance theory (ART) architecture, which models networks of neurons with feedback modulation. The next section reviews recent data on GWN development during infancy, focusing on the growth and subsequent activity-dependent pruning of long-range connections between the co-developing visuomotor, medial visual/categorization, attention and affective/reward networks. Perturbations of this process and their expected phenotypic presentations are then considered.

Development of the GNW

Phenomenology of the infant world

Like any developmental process, the development of the GNW can be expected to be sensitive to genetics, to the pre-, peri- and postnatal physical, chemical and biochemical environments, and to pre-, peri- and postnatal experience. As the GNW is a global functional network, its development can be expected to be particularly sensitive to the interplay between bottom-up effects originating in its “input” components, including proprioception and interoception, and top-down effects originating in more “central” components. The general Hebbian heuristic that synchronous activity enhances connectivity while asynchronous activity inhibits it can be expected to describe the functional integration of both the GNW itself and its component networks. At the level of individual cells, synchrony and asynchrony between even temporary synaptic partners can be expected to influence axonal branching and exploration, as recently demonstrated in non-mammalian systems (Munz et al. 2014; Kita et al. 2015).

Any successful model of GNW development must be consistent with the basic phenomenology of neonate and early-infant behavior. Human neonates are aware of and responsive to their environments and are particularly sensitive to motion, novelty, faces and sensations from within their own bodies (reviewed by Rochat 2012). They exhibit object-directed exploratory and aversive behaviors and rapidly develop gaze following, attention sharing and goal-directed object manipulation (reviewed by von Hofsten 2007). In the absence of other salient input, they provide themselves with sensory-motor feedback through motor and verbal babbling. The phenomenal world of the infant resolves itself only gradually, however, into the shared adult world of spatially-bounded, located, persistent objects with more or less stable collections of features (Colombo et al. 2012 provide a general review of infant cognition). Neonates can segregate moving objects from the background, for example, and are sensitive to features such as shape of static objects (Rakison and Yermoleva 2010). By 4 months old, infants can recognize moving objects as persistent following brief occlusion (e.g. Bremner et al. 2005). However, 4-month-olds may not segregate unfamiliar static objects from each other or from the background until their individual manipulability has been demonstrated (e.g. Needham et al. 2005). An ability to see features and an ability to see objects, in other words, does not automatically imply an ability to bind features to objects in a way that enables object segregation. Feature visibility does not, moreover, imply feature salience, which develops gradually and in a predictable order (e.g. shape and size before color) over the first year (reviewed by Baillargeon et al. 2011). The extent to which neonates and young infants experience segregated objects localized in space, as older infants, children and adults do, remains poorly understood (reviewed by Fields 2013).

The extent to which the early infant’s social world matches that of late infancy or childhood is also unclear. Both the feature clusters that indicate faces and the trajectory components that indicate animacy appear to be innate and available from birth (reviewed by Simion et al. 2011; Hoehl and Peykarjou 2012). Infants can identify and attribute goals to agents by as early as 3 months, with indicators of animacy serving as indicators of agency (reviewed by Luo and Baillargeon 2010; see also Luo 2011). It is, however, not clear whether young infants treat agency as a persistent feature of objects seen to act as agents, or whether “agent” is a category that generates specific expectations about behavior.

While parenting and other social interactions provide extrinsic motivations to correctly identify, categorize and associate expectations with objects, theoretical considerations (e.g. Friston 2010; Gottlieb et al. 2013), experiments using developmental robotics (reviewed by Oudeyer et al. 2013; Cangelosi and Schlesinger 2015) and observational and experimental studies of infant behavior (e.g. von Hofsten 2007; Gopnik 2012; Rochat 2012) all suggest that a powerful intrinsic motivation to increase the predictability of the phenomenal world drives this transition. It is important to bear in mind that both the particular static features and the particular motion characteristics of each of the identifiable objects comprising the infant’s world must be learned from observation. Infants must, for example, learn not just the specific facial features of family members and any others they recognize as familiar, but also their general visual features, voice, gait, other typical motions or gestures, smell, etc. These features, generated by multiple sensory pathways, must be integrated into a single object token (reviewed by Zimmer and Ecker 2010) maintained in memory as a stable, temporally-persistent individual that is permanently distinguished from all other individual objects. This representation must, moreover, be stable across periods of non-observation and against changes in location and minor changes in features. Infants must, for example, learn that changes in clothing—which may be highly visually salient—do not indicate changes in the identity of an individual person (reviewed by Fields 2012b); failures to recognize perceptual invariants that indicate object identity have, indeed, been proposed as an underlying functional correlate of ASD (Hellendoorn et al. 2015). Typically-developing infants achieve this level of object-token integration and stability for dozens if not hundreds of individual objects within the first year. There is every reason to believe that successful recognition of persistent objects generates significant affective rewards that encourage further exploration of both social and non-social components of the world.

Functional connectivity in the infant GNW

It has long been known that cortical functions develop in a roughly caudal-to-rostral direction, with sensory-motor networks maturing several years before prefrontal executive systems (reviewed by Casey et al. 2005). Several recent longitudinal resting-state fMRI or diffusion tensor MRI (DTI) studies of pre-term or full-term neonates, 1- and 2-year-olds (Doria et al. 2010; Yap et al. 2011; Gao et al. 2013; Alcauter et al. 2014; Ball et al. 2014; Gao et al. 2015; Huang et al. 2015; Toulmin et al. 2015; see also Di Martino et al. 2014; Dehaene-Lambertz and Spelke 2015; Vértes and Bullmore 2015 for supporting results) now allow following the early stages of large-scale functional network development at high resolution. These studies reveal robust connectivity within, and to varying extents between, major RSNs defined in adults (e.g. Damoiseaux et al. 2006; Yeo et al. 2011), including in particular the visuomotor, medial visual/categorization, dorsal and ventral attention, and affective/reward networks of primary interest here. Distinct RSNs including sensory, motor, attention and affective networks embedded in an SWN architecture with evident “rich club” nodes including frontal, parietal and superior temporal areas as well as hippocampus and amygdala are identifiable over 2 months pre-term (Doria et al. 2010; Ball et al. 2014). Adult-like cortico–cortico (Yap et al. 2011; Gao et al. 2015; Huang et al. 2015) and thalamo-cortical (Alcauter et al. 2014; Toulmin et al. 2015) RSN connection topology is evident in term neonates, with some departures from normal connectivity in early pre-term infants of an equivalent age (Doria et al. 2010; Ball et al. 2014).

The studies of Yap et al. (2011) and Gao et al. (2015) each examined neonates, 1- and 2-year-olds, using DTI and fMRI methods, respectively. On average, within-network connectivity increases rapidly during the first year and more slowly during the second year, with the exceptions of the sensory-motor network in which connectivity declines in both years and the language network which continues steady growth (Gao et al. 2015). Early postnatal increases in between-network connectivity are followed by connectivity decreases as between-network functional activity is refined by experience. As expected, sensory and sensory-motor networks develop earlier and faster than executive and default networks (see also Doria et al. 2010 for evidence that the same is true prenatally). The functional decoupling of the default and dorsal attention networks by 2 years (Gao et al. 2013) is particularly striking. The more rapid functional integration of frontal-parietal attention networks bilaterally in boys compared to girls demonstrated by Gao et al. (2015) is also particularly interesting given the atypical allocation of attention in ASD, the greater prevalence of ASD in males (e.g. Baron-Cohen 2002; Crespi and Badcock 2008), and recent demonstrations of reduced frontal-parietal connectivity at 12 months in high ASD-risk infants (Keehn et al. 2013; Righi et al. 2014; though see Uddin et al. 2013b; Picci et al. 2016 for cautions regarding the extrapolation of connectivity results across the lifespan).

While these studies of typically-developing neonates and infants do not explicitly discuss the experiential or behavioral correlates of their network-level observations, from a GNW perspective it is useful to consider the relation between increasing within- and between-network functionality and the phenomenology of perceiving and acting in the world. As noted earlier, neonates are capable of basic sensory-motor integration, consistent with the observation of Gao et al. (2015) that bilateral medial visual and sensory-motor networks have essentially adult-like connection topology at birth. Sensory-motor coordination increases rapidly as motor babbling gives way to more intentional and object-oriented activity; the comparatively early decreases in sensory-motor within-network connectivity observed by Gao et al. (2015) may reflect the early functional refinement of this system. The visuomotor (lateral visual/parietal in Gao et al. 2015) network is heavily involved in visual object segregation, motion tracking and the targeting and coordination of reaching behaviors; its continued growth during this period may reflect the increasing complexity of position, shape and motion information that infants are able to process as their behavioral capabilities mature. Early development of thalamo-cortical connections (Alcauter et al. 2014; Toulmin et al. 2015) is consistent with early emergence of multi-sensory integration as well as the robust awareness of their surroundings exhibited even by neonates. Robust connections between the reward (components of Salience in Gao et al. 2015), sensory, sensory-motor and attention networks are consistent with the assignment of strong affective valences to some objects—e.g. caregivers, strangers or favorite toys—as well as the presence of intrinsic motivations to explore the local environment and experiment with the objects it contains.

These behavioral and architectural observations, together with general neurodevelopmental considerations, suggest the qualitative picture of GNW development shown in Fig. 3. Regions within both cortex and subcortical areas are defined by some combination of genetic and spatial or other physical constraints during early embryogenesis. Within-region connectivity increases rapidly in the late 2nd to early 3rd trimester (reviewed by Andersen 2003). Long-range connectivity at this time is sparse and functionally incoherent (Fig. 3a). As the results of Doria et al. (2010) and Ball et al. (2014) indicate, both within-region and long-range connectivity increase significantly during the 3rd trimester, reaching sufficient long-range coherence to enable pre-term infants to exhibit object discrimination, exploratory behavior and social interactions (Fig. 3b). Regions with strong genetic determinants such as the fusiform face area or having early exposure to coherent input such as auditory cortex are sufficiently functional to enable individual recognition. By normal term, multi-regional functional networks are robustly connected, between-network connectivity has increased, and “rich club” hubs are apparent (Fig. 3c). By this stage, experiences processed primarily by one network can affect the perceptual expectations generated by a different network.

By the end of the first postnatal year, increases in connectivity are counterbalanced by experience-dependent synaptic pruning (Andersen 2003), as reflected in the slower growth in within-network connectivity during the second year observed for most RSNs by Gao et al. (2015). This process is effectively Hebbian and hence highly dependent on the coherence, consistency, and hence predictability of the experienced world. A higher probability of pruning for relatively short, within-region connections is indicated by the increases in longer within-network or between-network connections compared to shorter within-network connections during the second year noted by both Yap et al. (2011) and Gao et al. (2015).

A graphic notation for visualizing typical and variant GNW development

As shown by RSN imaging (Doria et al. 2010; Yap et al. 2011; Gao et al. 2013, 2015; Alcauter et al. 2014; Ball et al. 2014; Huang et al. 2015; Toulmin et al. 2015) and illustrated in Fig. 3, the connection probabilities within and between functionally-characterized regions are functions of both developmental time and the identities of the regions being connected. Even within-region connectivity increases and decreases at different times and rates in different regions, and both between-region connectivity within a given RSN and between-network connectivity vary significantly among individual RSNs and between pairs of RSNs, respectively. Within an RSN, the existence of a “rich club” of highly mutually-connected hubs even prenatally (Doria et al. 2010; Ball et al. 2014) indicates that connection probabilities exhibit preferential attachment (Barabasi and Albert 1999), a special case of assortative connectivity (Newman 2003) in which well-connected nodes have a higher probability of acquiring additional connections than poorly-connected nodes. Hebbian learning requires, moreover, that connection probabilities depend on synchrony of activity between new and existing connections. The probability P _ij(t) of a new connection between functional regions i and j at time t can, therefore, be expected to depend, at minimum, on genetically-established cell-surface or other markers identifying i and j as functional regions, the number of i–j connections already established, the correlation at t between the overall activities of regions i and j, and particularly for intra-regional (i.e. i = j) connections, the extent to which the new connection increases the temporal coherence of regional activity. While reciprocal connection probabilities may be expected to be similar (i.e. P _ij(t) ~ P _ji(t)), they need not be identical. Similar dependencies can be expected for connection probabilities between multi-regional RSNs, which themselves form a rich-club network (Sporns and Honey 2006; Shanahan 2012). It is difficult, given this level of complexity, to frame questions about the results of developmental shifts in the probabilities of particular region-to-region or network-to-network connections or about changes in the magnitudes of particular connection probabilities (e.g. Di Martino et al. 2014, Fig. 1) in this representation.

For the purposes of characterizing variants in GNW development, however, the time-dependent ratios P ^Var_ij(t)/P ^Typ_ij(t) of variant (Var) to typical (Typ) connection probabilities either within or between RSNs are more relevant than the connection probabilities themselves. Following Rubinov and Sporns (2010), we consider these probability ratios to incorporate information on the strength of connectivity as a function of time, i.e. we interpret P ^Var_ij(t)/P ^Typ_ij(t) as a measure of the time-dependent effective connectivity (Friston 2011) of a variant network compared to that of a typical network. As we do not assume that P _ij = P _ji for all i and j, these effective connectivity ratios are directional. Such ratios are amenable to a simple graphic representation, as shown in Fig. 4. At each fixed time t, the graph of directed connections between any given set of either N regions or N multi-region networks can be drawn as a complete directed graph (i.e. digraph) on N nodes, with symmetry under 2π/N rotations indicating typical connectivity (i.e. P ^Typ_ij(t)/P ^Typ_ij(t) = 1). Equal normalized effective connectivity between nodes is indicated by drawing equal-width connections. In graphs in which each node represents a network, e.g. an RSN, each of the N nodes is represented by superposed solid and open symbols, indicating that the ratio of short (i.e. within-region) connections in the network to long (i.e. between-region) connections within the network is typical. As “typical” connectivity is by definition typical at all developmental times, a typically-developing network will have the same graph for all t. With this graphic convention, any differences from the graph structure shown in Fig. 4a, including the unaligned solid and open node symbols, unequal connection widths and geometric asymmetries shown in Fig. 4b, represent differences from typicality.

Geometric distortions may be used to symbolize differences from typicality in the ratios between connection probabilities, as shown in Fig. 4b. Moving the open and solid symbols for each node apart indicates a larger (dot farther from center than circle) or smaller (circle farther from center than dot) ratio of short- to long-range connections within a network; moving both symbols outward or inward provides a representation for whole-network over- or under-development, respectively. Thicker lines between networks indicate higher-than-typical between-network connectivity; thinner (i.e. dashed) lines indicate lower-than-typical between-network connectivity. At such variations may occur at different developmental times, the graph of a variant network may have different geometry at different times.

In the next section, we consider the consequences that may be expected from early-developmental perturbations of GNW organization, focusing on the subnetwork of the GNW comprising the affective/reward, medial visual/categorization, visuo-motor, and dorsal and ventral attention systems. We then employ the graphic representation developed above to examine the functional disruptions proposed in recent second-generation models of ASD (Fields 2012a; Pellicano and Burr 2012; Lawson et al. 2014; Van de Cruys et al. 2014; Hellendoorn et al. 2015), showing that in each case they can be described as consequences of such connectivity perturbations.

Perturbations of GNW development

General versus localized developmental GNW disruptions

By placing the emergence of an SWN architecture linking functionally-specialized cortical and subcortical regions in the early third trimester at the latest, the results of Doria et al. (2010) and Ball et al. (2014), in particular, render the association of genes acting to regulate neural differentiation and connectivity in early- to mid-fetal development (Parikshak et al. 2013; Willsey et al. 2013) and maternal dietary deficiencies or toxin exposure (Currenti 2010; Gardener et al. 2011) with ASD less surprising. The results of Chow et al. (2012) indicate, moreover, that such genes continue to act through early childhood, as expected on the basis of gross anatomical and functional changes (Andersen 2003; Casey et al. 2005), but are replaced as potential markers of pathology by maintenance-associated genes in adulthood. Symptoms attributable to GNW disruption would, therefore, be expected to arise no later than early childhood, and to be detectable, at least as prodromes, either neo- or peri-natally.

Early neuron over-growth followed by delayed functional maturation and reduced white-matter connectivity, particularly in frontal and temporal cortex and amygdala, was the first well-documented neurological correlate of ASD (reviewed by Courchesne and Pierce 2005; Courchesne et al. 2005, 2007; for recent evidence that head size per se is not an information marker, see e.g. Zwaigenbaum et al. 2014). Uniform enhancement of short-range at the expense of long-range connectivity in these areas would be expected to disrupt function in all RSNs of interest here except the visuo-motor system, as well as in the language, default, and higher-level executive systems. Such extensive functional disruption would be expected to correlate with severe social, language, and attentional symptoms, consistent with the generally-recognized (e.g. Geschwind 2009) tendency to diagnose only severe cases as ASD until relatively recently. An over-responsive amygdala and dysregulated ventral attention system could, in particular, be expected to generate the “intense world” symptoms discussed by Markram et al. (2007), while general frontal hypofunctionality could explain the learning disabilities and general intelligence deficits of severe autism, as proposed in theories of autism that focus on executive dysfunction as the primary cognitive phenotype (Rajendran and Mitchell 2007).

The recent genetic dissociation of ASD and general intellectual disability (Parikshak et al. 2013), the varied symptomology of less-severe ASD and the diversity of ASD-associated genes and expression profiles all suggest that more localized and subtle disruptions of either regional or network-level function may be involved in most cases of ASD. While any significant departure from typicality in the GNW components shown in Fig. 4a, for example, can be expected to yield variant experience and behavior, the question of interest for pathology is whether such variation will be self-correcting in a typical infant environment, or will progress to diagnosable symptoms within the first 2–3 years. Using DTI and a connection-strength weighted efficiency measure, Lewis et al. (2014) have shown that RSN connection efficiency correlated with ASD severity at 2 years, noting that the networks most affected are those traversing ventral occipital and temporal cortex. This result indicates that at least some variants do not self-correct, among them variants associated with ASD. Relatively small variants in the pattern of connectivity that may be amplified, in a typical infant environment, into significant disruptions of function are of particular interest. Before discussing specific recent models of ASD, it is useful to review two kinds of variants that may be expected, on the basis of general neurocognitive considerations, to undergo such amplification: perturbations of salience and attention, and perturbations of category learning, including perturbations of category–reward associations. When viewed from the perspective of predictive coding and the Free Energy Principle (Friston 2010), these kinds of variants could be expected to affect expectation generation, error detection and coding, and precision estimates, as described in the models of Lawson et al. (2014) and Van de Cruys et al. (2014) discussed below. We also briefly review expected effects of such variants on the default and executive systems, as it is these consequential effects that would be expected to generate the deficits in theory of mind and central coherence typical of ASD.

Perturbations of salience

As noted earlier, motion, faces, bodily sensations and novelty are particularly salient to infants. The relatively high salience of novel stimuli provides the basis for the widely-used looking time paradigm in studies of infant cognition (reviewed by Aslin 2007). Most available data on the relative salience of different classes of visual features (e.g. shapes, sizes, colors) and events (e.g. apparent violations of object permanence or solidity, inanimate objects acting as agents) have been provided by looking-time studies (see Luo and Baillargeon 2010; Baillargeon et al. 2011; Aslin 2014 for relevant reviews); hence there is significant ambiguity between “salience” and “novelty” in the early-development literature. As expected from considerations of learning efficiency (e.g. Oudeyer et al. 2013), infants learn more effectively in the presence of novelty (Stahl and Feigenson 2015; see Gopnik 2012 for a review of supporting data), but divert attention away from overly-complex stimuli from which coherent regularities cannot be extracted (Kidd et al. 2012; see Aslin 2014 for a review of supporting data). Both intrinsic motivation and learning context are, therefore, also potential confounds in salience studies with infants. One can, however, ask how disruptions in the typical development of object and event salience would affect an infant’s experience of the world while taking these various caveats into account.

Both faces and motions indicative of animacy are key indicators of agency, and hence of the potential for social interactions, from earliest infancy onwards (Simion et al. 2011). Specific deficits in either face perception (e.g. under-connectivity of fusiform face area) or the detection of biological motion (e.g. under-connectivity of STS) would, therefore, be expected to disrupt the perception of agency and hence the ability to engage in and learn from social interactions. Deficit social interaction can be expected to impact social learning in general and language learning in particular; hence both face and biological motion perception deficits can be expected to amplify into generalized cognitive-affective deficits in a typical infant environment. Deficits in face recognition and the inference of emotion from facial expressions correlate with severity in ASD (reviewed by Harms et al. 2010; Weigelt et al. 2012), with considerable but not all evidence suggesting that emotional information conveyed by the eyes is specifically neglected (reviewed by Tanaka and Sung 2016). Deficits in coherent motion perception in general (Robertson et al. 2014) and biological motion perception in particular (reviewed by Kaiser and Pelphrey 2012) are associated with ASD, and are attributable to deficit higher-order processing, e.g. in STS, not deficit lower-level motion detection.

Salience controls the reactive ventral attention system and is controlled, at least in part, by the voluntary dorsal attention system (reviewed by Corbetta and Shulman 2002; Vossel et al. 2014; Uddin 2015). Disrupting the balance between these two competing systems in favor of reactivity, e.g. by over-activation of the amygdala in response to detected motion, would be expected to negatively impact attention sharing and hence social learning in addition to generating inappropriate negative affect. Disrupting the balance in favor of voluntary attention, e.g. by over-activation of motion-tracking areas in intraparietal sulcus, would be expected to negatively impact attention switching, again rendering attention sharing and social learning more difficult. Environments rich in unpredictable motion stimuli would be expected to amplify the former condition, while environments rich in predictable motion stimuli would amplify the latter. High distractability with inappropriate negative affect (e.g. Markram et al. 2007) and low distractability with deficit attention switching (e.g. Elsabbagh et al. 2013) are both described in ASD, suggesting that both of these mechanisms may contribute to ASD symptomology.

The existence of multiple, possibly opposing mechanisms capable of generating cognitive, affective and behavioral symptoms meeting diagnostic criteria for ASD may also explain apparently-conflicting connectivity results. Hyper-connectivity within the salience network, and specifically between insular and anterior cingulate cortices, has been demonstrated in pre-adolescent children with ASD (Uddin et al. 2013a, b). In a study of primarily older children and adolescents, however, Abbott et al. (2015) demonstrated hypo-connectivity between insular and anterior cingulate cortices. A similar situation exists regarding the potential role in ASD of von Economo neurons (VENs), large spindle-shaped cells with long projections found in both insular and cingulate cortices. Postmortem studies of small numbers of subjects have reported fewer (Simms et al. 2009, 6 of 9 subjects), the same number (Kennedy et al. 2007) and more (Simms et al. 2009, 3 of 9 subjects; Santos et al. 2011) VENs relative to pyramidal cells in ASD subjects that in controls (see Allman et al. 2011; Butti et al. 2013 for reviews). As in the case of general salience network connectivity, the apparent conflict between these results may be resolvable by considering a broader functional context that takes the potential effects of small local changes on other networks into account. Such “knock-on” effects determine what kinds of internal or external stimuli have reduced or enhanced salience, and hence determine the cognitive, affective and behavioral consequences of local variants in neuron numbers or connectivity. Over a developmental time scale, small differences in the response of other networks to increases or decreases in salience can lead to dramatically different outcomes; for example, whether an increase in the salience of motion stimuli is associated with an increased fear response may determine the difference between an “intense world” (Markram et al. 2007) outcome and a more benign hypersystemizing (Fields 2012a) outcome.

Perturbations of categorization

Again as noted earlier, the grouping of objects into categories with which appropriate expectations can be associated is a key requirement for achieving predictability of and some level of control over the environment. It is also a pre-requisite for language learning (reviewed by Westermann and Mareschal 2012). In adults, object categories or “concepts” are implemented as associations between modally-represented object features and behaviors (reviewed by Kiefer and Pulvermüller 2012; see also Fernandino et al. 2015) and typically point to multiple episodic-memory resident object tokens representing individual exemplars (Fields 2012b). The implementation of categories in infants is not well characterized, even in the canonical case of human faces (reviewed by Hoehl 2016). The results of Gao et al. (2015) indicate that connectivity of the medial-visual network is adult-like in neonates; however, they provide no information on higher-order categorization areas (e.g. anterior temporal pole) or processes (e.g. abstraction or the organization of inter-category relations). The rapid development of categorization capabilities after language capabilities have developed suggests that categorization may be not just rudimentary but possibly also fluid in pre-linguistic infants.

At least four kinds of categorization disruption can be expected to be amplified into significant dysfunction in a typical infant environment. First, categorizing objects correctly whether they are currently moving or not requires the suppression of motion information relative to feature information during both category learning and subsequent object categorization (Fields 2011). Insufficient suppression of motion information during category learning can be expected to result in aberrant, motion-emphasizing or even motion-dominated categories, with deleterious consequences for object identification and language learning (Fields 2012a; see also below). Second, the animate-inanimate distinction is fundamental to normal category formation and may indeed be innate (Simion et al. 2011). Disruptions of this distinction can be expected to result in aberrant agency judgments and either too-broad or too-narrow social interactions. Third, many categories incorporate affective valences; either enhancing or suppressing interactions between the categorization and reward systems can be expected to lead to inappropriate approach or avoidance behaviors and either over- or under-expression of affect. Finally, many categories incorporate manipulability information or even specific motor plans. Suppression of such connections would be expected to lead to category-specific apraxias as noted earlier, while enhancements may lead to inappropriate attempts to manipulate objects or styles of manipulation. Motor-planning deficits and dyspraxias are common in ASD (Jeste 2011), although a connection with categorization deficits has not been established. Inappropriate manipulative and other motor behavior is also common in ASD.

Deficit categorization or the deployment of robust but aberrant categories can be expected to lead to the generation of either deficit or aberrant expectations about the behavior of objects in the world, including other people. Considerable evidence supports an association between deficit or aberrant expectation generation and hence deficit or aberrant action planning with ASD (reviewed by Gomot and Wicker 2012; see also discussions of Pellicano and Burr 2012 and Van de Cruys et al. 2014 below). Deficits or significant variants in the detection and processing of salience as discussed above would, moreover, be expected to generate or at least correlate with deficit or variant categorization. The medial-visual/categorization network (MV/C in Fig. 4) comprising the static-feature detecting ventral visual processing stream, feature-cluster detectors in fusiform and lingual gyrus, hippocampal and parahippocampal areas associated with binding, object recognition and episodic-memory encoding, and high-level “concept” areas in the anterior temporal pole would be expected to be “ground zero” for such deficits or variants. The network connection inefficiencies correlated with ASD in 2-year-olds by Lewis et al. (2014) are, interestingly, concentrated in these areas.

Consequences of salience and categorization perturbations for the default and executive systems

As noted earlier, the executive and default networks develop later and mature more slowly than the sensory processing and sensory-motor control networks of primary interest here. From a cognitive-behavioral standpoint, major milestones in the development of these networks include, respectively, reasoning about mechanical causation and theory-of-mind (ToM), both of which are typically achieved in the pre-school years (reviewed by Fields 2014), and abstract (e.g. mathematical or mechanical) and personally-relevant planning, which are typically achieved by adolescence (Casey et al. 2005; for the default network in particular, Fair et al. 2008). First-generation models of ASD centered around deficits in central coherence (e.g. Happé and Frith 2006) or ToM (e.g. Baron-Cohen 2002) focused attention on deficits in or disruptions of the executive or default networks, respectively, while functional and connectivity studies supporting disconnection of prefrontal cortex in ASD relative to typical development (e.g. Courchesne and Pierce 2005; Geschwind and Levitt 2007; Wass 2011) suggested an etiological focus on these networks. However, recent studies suggest a more complicated picture. Shih et al. (2010) and Abbott et al. (2015), for example, both report overconnectivity between executive and default networks together with underconnectivity between specific components of each network; Elton et al. (2016) similarly report enhanced default-to-executive connectivity combined with disrupted connectivity between components of the executive network. Such results are consistent with the more nuanced view of connectivity variations in ASD that has emerged as larger subject populations and refined methods have become available (Müller et al. 2011; Vissers et al. 2012; Maximo et al. 2014; Picci et al. 2016).

Within the current, GNW-based framework, variant functional connectivity within and between the executive and default networks, as well as between these networks and the categorization, salience and attention systems, is expected to result, at least in large part, from disrupted or variant functional connectivity within and between earlier-developing networks. Disruptions of ToM or self-representation, for example, can be expected to result from disruptions of the categorization and salience systems, as discussed further below. A categorization network biased toward similarities in kinematic or dynamic properties of objects and against similarities in static properties would, similarly, be expected to facilitate planning and abstract problem solving in formal and mechanical domains while inhibiting planning and problem solving in the social domain (Fields 2012a, b). Substantial individual variation in executive and default network functions are, moreover, to be expected if disruptions in these networks are developmental consequences of disruptions or variations in functionally antecedent networks with which these later-developing systems have bidirectional connectivity. Such a scenario is consistent with both the heterogeneity of connectivity variations (Hahamy et al. 2015) and symptomatic presentations observed in ASD adults.

Five recent models of ASD from a GNW perspective

As noted earlier, disruptions of functional connectivity in ASD have been intensively investigated since the early 2000s (Courchesne and Pierce 2005; Geschwind and Levitt 2007; Minshew and Williams 2007; Rippon et al. 2007; Müller et al. 2011; Wass 2011; Vissers et al. 2012; Maximo et al. 2014; Tyszka et al. 2014; Picci et al. 2016) and ASD is now widely considered to be a connectome-scale developmental disorder (Di Martino et al. 2014; Dehaene-Lambertz and Spelke 2015; Vértes and Bullmore 2015). While the vast majority of studies have examined connectivity in older children, adolescents or adults with ASD, significant departures for typical connectivity have been observed, albeit at relatively low spatial resolution, even in high ASD-risk infants (Keehn et al. 2013; Righi et al. 2014). As the GNW is the “network of networks” that coordinates conscious activity (Dehaene and Changeux 2011; Dehaene et al. 2014), it is reasonable to ask how changes in connectivity might affect GNW function and hence experience of and interaction with the world. It is, in particular, reasonable to ask how such changes might affect the functions of the early-developing sub-network of the GNW represented in Fig. 4a.

Even second-generation, neurocognitive models of ASD have focused on mechanisms that might underlie typical cognitive and behavioral manifestations of ASD post-diagnosis, i.e. beginning in early childhood. The considerations outlined in the previous two sections suggest, however, that post-diagnosis symptoms may have their source in very early, possibly prenatal alterations in functional connectivity. Our goals in this section are (1) to show that the neurocognitive outcomes proposed by the models of Fields (2012a), Pellicano and Burr (2012), Lawson et al. (2014), Van de Cruys et al. (2014) and Hellendoorn et al. (2015) can be represented within the graphical framework of Fig. 4 and (2) to suggest that in every case, these outcomes may result from relatively small, localized, and early perturbations of connectivity within the medial-visual/categorization network. We review each model briefly, outlining in particular the symptoms or typical characteristics of ASD it seeks to explain; we then reformulate each model in the graphical framework of Fig. 4 and use this reformulation to highlight connectivity perturbations that could produce the initial conditions assumed by the model.

As noted earlier, perturbations of MV/C network can be expected to manifest as disruptions of categorization. Early connectivity perturbations of this or other networks are consistent with the mid-gestation action of many ASD-associated genes (Parikshak et al. 2013; Willsey et al. 2013). As the developmental consequences of small, localized and early connectivity perturbations can be expected to be highly dependent on interactions both with unperturbed components of the neurocognitive network and with the external environment, including the prenatal environment, considerable variation in phenotypic presentation by the age of typical diagnosis can be expected within this GNW-based framework.

Motion-based categories (Fields 2012a)

Fields (2012a) suggested that insufficient suppression of inputs from the dorsal visual processing (i.e. visuo-motor) stream relative to inputs from the ventral (i.e. static feature recognition) stream during category learning could lead to the generation of categories that grouped objects by their motion patterns, not by their static features. Such motion-based categories would, in the extreme case, be orthogonal to typical human feature-based categories; a family member, for example, might be assigned to different categories depending on how he or she was moving at the time of observation, and hence might never be recognized to be a single, situation-invariant individual. Fields (2012a) argued that deploying such motion-based categories instead of feature-based categories in infancy would broadly interfere with typical object-directed behaviors, social interactions and language learning, leading to cognitive and behavioral symptoms typical of ASD.

Both the initial perturbation postulated by this model and its amplified outcome are represented in Fig. 5 using the visualization scheme developed earlier (cf. Fig. 4). The initial perturbation is an increase in input connectivity from the visuo-motor network to the medial visual/categorization network, with an accompanying decrease in feature- or category-driven inhibition of motion information. Continued activation of motion-based categories would be expected to further decrease the effective contribution of static features to the categorization process, and hence to further decrease effective connectivity from the medial visual system to higher-level medial and anterior temporal areas involved in category learning and memory. Enhanced correlation between visuo-motor and category inputs to other networks together with decreased correlation between medial visual (i.e. feature-encoding) inputs and category inputs to these same networks would tend to increase outgoing visuo-motor connectivity while decreasing outgoing medial visual connectivity as indicated in Fig. 5. As objects would be recognized primarily by their motion patterns, dorsal attention system interactions with the visuo-motor system would be strengthened.

While they are broadly consistent with ASD phenomenology, the connectivity changes predicted in Fig. 5 have not been directly tested. Enhanced visuo-motor connectivity relative to medial visual connectivity is, however, suggested by results of Zmigrod et al. (2013), who shows that older ASD children were both faster than typical children in a spatial response task and less accurate than typical children in adjusting a spatial response given a featural (shape or color) cue. Using an object identification and location task monitored with fMRI, DeRamus et al. (2014) showed both that older ASD children are marginally deficient in object identification but not object location compared to typicals, and that temporal–parietal functional connectivity is reduced in ASD children relative to typicals, in both cases consistent with the model.

Attenuated expectations (Pellicano and Burr 2012)

Pellicano and Burr (2012) introduced the idea that an overall attenuation of top-down expectations could explain the enhanced sensitivity to perceptual detail, decreased sensitivity to gestalt features of scenes, and variant emotional responses typical of ASD. When formulated in the terms of a Bayesian analysis of perception as an inverse inference of features of a stimulus from features of its image, top-down expectations become experience-based prior probabilities of stimulus features; a uniform prior probability distribution corresponds to no expectations about what is contained in a scene (e.g. Knill and Pouget 2004). Reducing the relative strength of expectations by flattening the prior probability distribution allows image features to determine the perceived stimulus features with minimal expectation-based interpretation. Pellicano and Burr (2012) suggest that relatively flat prior probability distributions for most stimulus features would lead to excessively “literal” and detail-oriented perceptions, relative immunity to perceptual illusions, preference for exactly-repeated stimuli, and aversion to perceptual noise and rapid changes of scene.

As shown by Bastos et al. (2012) and Adams et al. (2015), fine-scale predictions about typical activity patterns and hence prior probabilities in the Bayesian sense are encoded even by the local-circuit architecture of cortical minicolumns. In principle, therefore, all cortical processing can be viewed as Bayesian inference. For the present purposes, however, it is useful to focus on the encoding of prior probabilities for perceived scenes by the categorization system. Category learning during infancy and childhood can be viewed as the progressive elaboration of experience-based prior probabilities for perceptible features within each category (e.g. Gopnik and Wellman 2012). Developing the expectation that dogs have fur, for example, is incorporating the feature `has-fur’ into the category `dogs’ or, in Bayesian terms, increasing the prior probability P(fur) for objects already classified as dogs or, more generally, increasing the prior conditional probability P(fur|dog). Relatively uniform probability distributions for features correspond, in this case, to relatively open-ended and uninformative categories. Extending the relatively uniform probability distribution for the number of limbs appropriate to the `animals’ category to the subcategory `mammals’, for example, leads to a significant decrease in predictive power; not including as an essential feature that mammals have faces similarly decreases the power of the category. Unlike in the model of Fields (2012a), the attenuation of prior probabilities and hence perceptual expectations proposed by Pellicano and Burr (2012) applies equally to all features, including features specifying motion patterns. The model of Pellicano and Burr (2012) does not predict that individuals with ASD will categorize individuals by their motion patterns, but rather that motion patterns, like static features, will be incorporated only weakly into categories.

Encoding a rich category structure incorporating probabilities for large numbers of features within each category as well as conditional probabilities relating features to each other requires high coding capacity and hence a wealth of short-range connections within the categorization system. The general attenuation of perceptual expectations and hence of categorization proposed by Pellicano and Burr (2012) could be expected, therefore, to result from an early-developmental perturbation of GNW architecture that decreased short-range connectivity and hence information encoding capacity within the medial-to-anterior temporal-lobe network that implements the categorization system, as shown in Fig. 6. Decreased top-down categorization feedback would be expected to decrease both top-down and bottom-up connectivity between both the dorsal and ventral visual processing streams and the categorization system, with concomitant decreases in ventral-stream connectivity to the attention and affective/reward systems. Attention and affective feedback would, in this case, be expected to be captured mainly by motion stimuli via the fast, categorization-independent, visuo-motor to limbic system pathways that enable reflexive responses to dangerous stimuli prior to conscious perception. Sensitivity to perceptual noise and change would be expected to be more severe on this model, for a given level of static-feature based categorization dysfunction, than on the model of Fields (2012a). However, the inappropriate behaviors based on inappropriate motion-based categories, i.e. behaving in the same way toward all entities that moved in the same way, expected from the model of Fields (2012a) would not be expected on this model. Pellicano and Burr (2012) make no specific claims about the severity of attenuation of expectations in their model, consistent with the graded dysfunctions seen in ASD.

As seen in Fig. 6, the long-term effects that could be expected to result from the attenuation of expectations proposed by Pellicano and Burr (2012) are not dissimilar to those predicted by Fields (2012a). Both models, in particular, predict over-development or over-activation of connections between the visuo-motor and dorsal attention system in ASD compared to typical development. The models are, however, distinguished by the deficit local connectivity within the categorization system predicted by the above reconstruction of the model of Pellicano and Burr (2012). While this prediction has not been tested directly, Keehn et al. (2013) reported marginally enhanced left-hemisphere anterior-to-posterior connectivity in high compared to low ASD risk 3 month-old infants, followed by decreased intrahemispheric connectivity in high-risk infants at 12 months; Righi et al. (2014) similarly reported decreased intrahemispheric connectivity in high-risk infants at 12 months. Additional support for reduced temporal-lobe connectivity in late infancy is provided by Lombardo et al. (2015), who demonstrated reductions in temporal lobe activity in response to speech in 1- to 4-year-olds with ASD and poor language development compared to age-matched controls or ASD with good language development, and by Dinstein et al. (2011), who demonstrated reduced interhemispheric synchronization in superior temporal gyrus, an area implicated in language processing, in 1- to 3.5-year-olds with ASD compared to controls.

Overly-precise perceptions (Lawson et al. 2014; Van de Cruys et al. 2014)

As noted earlier, considerable evidence supports the ubiquitous use of predictive coding in mammalian cortical processing. In a predictive coding context, “correct” top-down expectation signals are insufficient to assure correct processing of perceptual inputs; the process of comparing perceptual inputs with expectations must also work correctly. Lawson et al. (2014) and Van de Cruys et al. (2014) suggest that it is not an attenuation of expectations per se that generates the typical perceptual and categorization symptoms of ASD, but rather a dysfunction in the process of generating the “error” signals that represent differences between perceptions and expectations. The magnitude of the error signal depends not only on the strengths but also on the estimated precisions of both the bottom-up perceptual signal and the top-down expectation or prior-probability signal; imprecise percepts are easier to match to expectations and vice versa, while very precise percepts are harder to match to expectations and vice versa. Precision is itself context-dependent and must be estimated based on experience; one has to learn, for example, when to regard either percepts or expectations with near certainty and when to be open to the possibility that one’s expectations are only weakly supported by insufficient knowledge or one’s visual interpretation of a scene is incorrect. As Friston et al. (2012) point out in a Commentary on Pellicano and Burr (2012), estimating precision is effectively a metacognitive process, as it requires inferences about the functional adequacy of already-complex cognitive representations.

The models of Lawson et al. (2014) and Van de Cruys et al. (2014) both focus on disrupted estimates of precision, but differ in the specificity of the assumed dysfunction. Lawson et al. (2014) suggest that in ASD sensory precision is in general set too high, while pointing out that “overly precise estimates of sensory precision and under-precise estimates of prior precision would produce the same functional consequences; i.e., perception/interaction that lies closer to the sensory input and is insensitive to context” (p. 6). Insufficiently-precise priors are, on this model, an expected side-effect (cf. van Boxtel and Lu 2013). Van de Cruys et al. (2014) locate the dysfunction not in the representation of percepts or priors per se, but in the atypical weighting of prediction errors and in the precision assigned to the percept-expectation error signal. This “high, inflexible precision of prediction errors in autism” (p. 4) or HIPPEA model is, at least in principle, independent of the precisions assigned to either percepts or expectations individually. Its emphasis on inflexibility, moreover, suggests that the assumed dysfunction cannot be overcome by learning.

Like the model of Pellicano and Burr (2012), the models of Lawson et al. (2014) and Van de Cruys et al. (2014) are intended to characterize ASD in children, adolescents and adults, i.e. in neurocognitive systems that have already developed robust, hierarchical categories. For the present purposes, the question of interest is how a categorization system characterized by overly-precise percepts and weak, imprecise priors or by inflexibility in precision assignments could develop during early infancy. Like perceptual expectations themselves, appropriate estimates of both perceptual and categorization precision must be learned from experience. Connectivity disruptions that impair such learning can be expected to impair estimates of precision. An insufficiently-developed categorization system, as shown in Fig. 6, would be expected to produce poorly-defined and hence imprecise expectations across the board. If all expectations are equally imprecise, accurate precision estimates cannot be learned. Alternatively, a failure to develop sufficient top-down connectivity from a normally-developing categorization system to normally-developing dorsal and ventral visual-processing streams, as shown in Fig. 7, could also be expected to disrupt precision estimates and hence precision-estimate learning. As insufficient or inaccurate feed-forward error signals would be expected to further disrupt category learning, the outcome of this perturbation could be essentially indistinguishable from the outcome of the Pellicano and Burr (2012) mechanism shown in Fig. 6.

Fig. 7 — Deficit top-down feedback of expectations to visual processing streams would be expected to result in categorization deficits similar to those expected from deficit categorization-system development (cf. Fig. 6)

The possibility that multiple initial perturbations of GNW development could produce functionally equivalent or at least very similar outcomes after exposure to typical infant and early-childhood environments is consistent with the idea that ASD is a syndrome with possibly quite-diverse etiologies. Disrupted connectivity within the MV/C system or between this system and the functionally very closely related VM system (i.e. the dorsal visual processing stream) appear, however, to be plausible candidates for the initial disruption across a broad range of models. It is interesting in this regard that connection efficiency within lingual gyrus, an MV/C component involved in complex-feature detection and binding, is both highly sensitive to prenatal genetic variability (Fornito et al. 2011) and highly correlated with ASD as an outcome (Lewis et al. 2014). These results suggest that lingual gyrus may be a plausible locus for the initial disruptions of MV/C connectivity represented in Figs. 5, 6 and 7.

Disrupted invariance detection (Hellendoorn et al. 2015)

Hellendoorn et al. (2015) offer a variation on the model of Pellicano and Burr (2012) in which what is attenuated in ASD is not perceptual expectations per se but the detection of perceptual invariants, where these are viewed in Gibson’s (1979) terms as affordance-specifying properties of the environment, including (for vision) the ambient photon field. The core of the model is the proposal that categorization in ASD is based on comparisons between percepts and individual exemplars as opposed to abstracted categories; contrasting their model with that of Pellicano and Burr (2012), Hellendoorn et al. (2015) “do not suggest that people with ASD have weak priors, but different priors. Because of their invariance detection impairments, we hypothesize that the priors of people with ASD also include variant aspects of the environment and are more exemplar-based instead of prototype-based” (p. 9, emphasis in original). People with ASD would, in other words, expect to encounter the same individual entity in the same environmental context, as opposed to expecting a “typical” entity in a “typical” context. From a developmental perspective, one would expect purely exemplar-based categorization in childhood or adulthood to result from a systematic failure of abstraction in infancy and hence from early dysfunction in the categorization system, although how such a process would be described in Gibsonian terms, which require all information about category membership to be encoded by the stimulus, is unclear.

As Hellendoorn et al. (2015) point out, expectations and hence priors based on individual exemplars cannot distinguish invariant from accidental and hence variant features of a category member; hence the substitution of exemplars for abstract categories would be expected to impair invariance detection. However, it is unclear how exemplars are associated with categories—in the example employed by Hellendoorn et al. (2015), how exemplars of factories are represented as exemplars of a common category `factory’—in the model, especially given the Gibsonian rejection of representation as a function of cognitive systems. However this association is implemented, however, it must be learned; hence the deficit in invariance detection postulated by Hellendoorn et al. (2015) may result developmentally from a deficit in invariance learning. The model of Hellendoorn et al. (2015) can, therefore, be reconstructed along the lines of either Figs. 6 or 7, i.e. as an outcome of either deficit categorization per se or of deficit category learning.

Conclusion

The second-generation models briefly reviewed here are all consistent with the emerging idea that “sensory processing is not only an additional piece of the puzzle, but rather a critical cornerstone for characterizing and understanding ASD” (Baum et al. 2015, p. 140). As discussed in detail by Fields (2012a), Pellicano and Burr (2012), Lawson et al. (2014), Van de Cruys et al. (2014) and Hellendoorn et al. (2015), these perception and categorization focused models are able to explain the deficits in social interaction, affect and executive function on which the earlier generation of models tended to focus as consequences of disruptions in the processing and interpretation, in context, of perceptual input. What has been suggested here is that the specific sensory processing, or in the case of Hellendoorn et al. (2015), sensory detection deficits or variants proposed these models can be understood as developmental outcomes of perturbations of GNW organization and function in early infancy, possibly prenatally. In particular, both the models considered here individually and the integrative framework we have proposed to understand them from a GNW perspective locate the etiology of ASD in the interactions between early-developing sensory processing, sensory-motor and attention systems, and regard functional disruptions in the “higher” executive and default networks as consequences, not causes. An early-developmental etiology of ASD is consistent with the early activity of ASD-associated genes (Parikshak et al. 2013; Willsey et al. 2013; see also Geschwind and Flint 2015) as well as the association between ASD and prenatal toxicity (Currenti 2010; Gardener et al. 2011). Functional disruption of GNW development is also consistent with the emerging view that ASD is a connectome-level developmental disorder (Di Martino et al. 2014; Dehaene-Lambertz and Spelke 2015; Vértes and Bullmore 2015).

The models discussed here have yet to be directly tested by high-resolution connectivity studies in infants, and they are supported at present only by low-resolution, indirect or circumstantial evidence as discussed above and in their respective original publications. High-resolution functional connectivity analysis of infant ASD (or high ASD risk) populations compared to typical controls, especially of pre-term infants followed longitudinally as performed by Doria et al. (2010) and Ball et al. (2014), may in principle be capable of distinguishing the mechanisms proposed by these models. As noted earlier, underconnectivity or functional disruptions in lingual gyrus may be an attractive target for such studies. At least two kinds of formal studies may also contribute. First, the developmental robotics methods employed by Schlesinger et al. (2012) to replicate typical infant performance on a perceptual completion task could, in principle, be extended to investigate the representational and algorithmic assumptions necessary to replicate typical and ASD infant performance on perceptual categorization tasks (see also Cangelosi and Schlesinger 2015 for a more general discussion of such methods). Second, network-growth methods such as those of Jarman et al. (2014), if extended to include position-dependent connection probabilities between functional regions as discussed above, may be sufficiently powerful to model the early development of the GNW in a way that provides new insights into network parameters that may be affected by genes, particularly axonal growth-cone guidance genes (e.g. McFadden and Minshew 2013), acting in mid- to late-fetal development.

Setting the differences between these models aside, however, they all suggest that early-developing deficits or disruption in the medial-visual/categorization network underlie ASD. They therefore suggest that therapeutic interventions that facilitate learning both feature-based categories and motion-feature correlations in early infancy may be useful for ameliorating the functional deficits of ASD.

Acknowledgements

We acknowledge, with thanks, comments from Dr. Sander van de Cruys and from two anonymous referees.

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflicts of interest relevant to the reported research.

Contributor Information

Chris Fields, Email: fieldsres@gmail.com.

James F. Glazebrook, Email: jfglazebrook@eiu.edu

References

Abbott AE, Nair A, Keown CL, Datko M, Jahedi A, Fishman I, Müller R-A. Patterns of atypical functional connectivity and behavioral links in autism differ between default, salience, and executive networks. Cereb Cortex. 2015;26:4034–4045. doi: 10.1093/cercor/bhv191. [DOI] [PMC free article] [PubMed] [Google Scholar]
Achard S, Salvador R, Whitcher B, Suckling J, Bullmore E. A resilient, low-frequency, small-world, human brain functional network with highly connected association cortical hubs. J Neurosci. 2006;26:63–72. doi: 10.1523/JNEUROSCI.3874-05.2006. [DOI] [PMC free article] [PubMed] [Google Scholar]
Adams RA, Friston KJ, Bastos AM. Active inference, predictive coding and cortical architecture. In: Casanova MF, Opris I, editors. Recent advances in the modular organization of the cortex. Berlin: Springer; 2015. pp. 97–121. [Google Scholar]
Alcauter S, Lin W, Smith JK, Short SJ, Goldman BD, Reznick JS, Gilmore JH, Gao W. Development of thalamocortical connectivity during infancy and its cognitive correlations. J Neurosci. 2014;34:9067–9075. doi: 10.1523/JNEUROSCI.0796-14.2014. [DOI] [PMC free article] [PubMed] [Google Scholar]
Allman JM, Tetreault NA, Hakeem AY, Manaye KF, Semendeferi K, Erwin JM, Park S, Goubert V, Hof PR. The von Economo neurons in fronto-insular and anterior cingulate cortex. Ann N Y Acad Sci. 2011;1225:59–71. doi: 10.1111/j.1749-6632.2011.06011.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
American Psychiatric Association . Diagnostic and statistical manual of mental disorders (DSM-V) Arlington, VA: American Psychiatric Association; 2013. [Google Scholar]
Andersen SL. Trajectories of brain development: point of vulnerability or window of opportunity? Neurosci Biobehav Rev. 2003;27:3–18. doi: 10.1016/S0149-7634(03)00005-8. [DOI] [PubMed] [Google Scholar]
Andrews-Hanna JR, Smallwood J, Spreng RN. The default network and self-generated thought: component processes, dynamic control, and clinical relevance. Ann N Y Acad Sci. 2014;1316:29–52. doi: 10.1111/nyas.12360. [DOI] [PMC free article] [PubMed] [Google Scholar]
Aslin RN. What’s in a look? Dev Sci. 2007;10:48–53. doi: 10.1111/j.1467-7687.2007.00563.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
Aslin RN. Infant learning: historical, conceptual, and methodological challenges. Infancy. 2014;19:2–27. doi: 10.1111/infa.12036. [DOI] [PMC free article] [PubMed] [Google Scholar]
Baars B. A cognitive theory of consciousness. New York: Cambridge University Press; 1998. [Google Scholar]
Baars BJ. Global workspace theory of consciousness: toward a cognitive neuroscience of human experience. Prog Brain Res. 2005;150:45–53. doi: 10.1016/S0079-6123(05)50004-9. [DOI] [PubMed] [Google Scholar]
Baars BJ, Franklin S. How conscious experience and working memory interact. Trends Cogn Sci. 2003;7:166–172. doi: 10.1016/S1364-6613(03)00056-1. [DOI] [PubMed] [Google Scholar]
Baars BJ, Franklin S, Ramsoy TZ. Global workspace dynamics: cortical “binding and propagation” enables conscious contents. Front Psychol. 2013;4:200. doi: 10.3389/fpsyg.2013.00200. [DOI] [PMC free article] [PubMed] [Google Scholar]
Baillargeon R, Li J, Gertner Y, Wu D. How do infants reason about physical events? In: Goswami U, editor. The Wiley-Blackwell handbook of child cognitive development. 2. Oxford: Blackwell; 2011. pp. 11–48. [Google Scholar]
Ball G, Alijabar P, Zebari S, Tusor N, Arichi T, Merchant N, Robinson EC, Ogundipe E, Ruekert D, Edwards AD, Counsell SJ. Rich-club organization of the newborn human brain. Proc Natl Acad Sci USA. 2014;111:7456–7461. doi: 10.1073/pnas.1324118111. [DOI] [PMC free article] [PubMed] [Google Scholar]
Barabasi AL, Albert R. Emergence of scaling in random networks. Science. 1999;286:509–512. doi: 10.1126/science.286.5439.509. [DOI] [PubMed] [Google Scholar]
Baron-Cohen S. The extreme male brain theory of autism. Trends Cogn Sci. 2002;2:248–254. doi: 10.1016/S1364-6613(02)01904-6. [DOI] [PubMed] [Google Scholar]
Bassett DS, Bullmore E. Small world brain networks. Neurosci. 2006;12:512–523. doi: 10.1177/1073858406293182. [DOI] [PubMed] [Google Scholar]
Bastos AM, Usrey WM, Adams RA, Mangun GR, Fries P, Friston KJ. Canonical microcircuits for predictive coding. Neuron. 2012;76:695–711. doi: 10.1016/j.neuron.2012.10.038. [DOI] [PMC free article] [PubMed] [Google Scholar]
Baum SH, Stevenson RA, Wallace MT. Behavioral, perceptual, and neural alterations in sensory and multisensory function in autism spectrum disorder. Prog Neurobiol. 2015;134:140–160. doi: 10.1016/j.pneurobio.2015.09.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
Belmonte MK, Baron-Cohen S (2004) Small-world network properties and the emergence of social cognition: evidence from functional studies of autism. In: Triesch J, Jebara T (eds) Proceedings of the 2004 international conference on development and learning. UCSD Institute for Neural Computation, La Jolla, CA, p 268
Betancur C. Etiological heterogeneity in autism spectrum disorders: more than 100 genetic and genomic disorders and still counting. Brain Res. 2011;1380:42–77. doi: 10.1016/j.brainres.2010.11.078. [DOI] [PubMed] [Google Scholar]
Block N, Carmel D, Fkeming SM, Kentridge RW, Koch C, Lamme VAF, Lau H, Rosenthal D. Consciousness science: real progress and lingering misconceptions. Trends Cogn Sci. 2014;18:556–557. doi: 10.1016/j.tics.2014.09.004. [DOI] [PubMed] [Google Scholar]
Brandi M-L, Wohlschläger A, Sorg C, Hermsdörfer J. The neural correlates of planning and executing actual tool use. J Neurosci. 2014;34:13183–13194. doi: 10.1523/JNEUROSCI.0597-14.2014. [DOI] [PMC free article] [PubMed] [Google Scholar]
Bremner JG, Johnson SP, Slater A, Mason U, Foster K, Cheshire A, Spring J. Conditions for young infants’ perception of object trajectories. Child Dev. 2005;76:1029–1043. doi: 10.1111/j.1467-8624.2005.00895.x. [DOI] [PubMed] [Google Scholar]
Butti C, Santos M, Uppal N, Hof PR. Von Economo neurons: clinical and evolutionary perspectives. Cortex. 2013;49:312–326. doi: 10.1016/j.cortex.2011.10.004. [DOI] [PubMed] [Google Scholar]
Buxbaum LJ, Shapiro AD, Coslett HB. Critical brain regions for tool-related and imitative actions: a componential analysis. Brain. 2014;137:1971–1985. doi: 10.1093/brain/awu111. [DOI] [PMC free article] [PubMed] [Google Scholar]
Cangelosi A, Schlesinger M. Developmental robotics: from babies to robots. Cambridge, MA: MIT Press; 2015. [Google Scholar]
Casey BJ, Tottenham N, Liston C, Durstan S. Imaging the developing brain: what have we learned about cognitive development? Trends Cogn Sci. 2005;9:104–110. doi: 10.1016/j.tics.2005.01.011. [DOI] [PubMed] [Google Scholar]
Chow ML, Pramparo T, Winn ME, Barnes CC, Li H-R, Weiss L, Fan J-B, Murray S, April C, Belinson H, Fu X-D, Wynshaw-Boris A, Schork NJ, Courchesne E. Age-dependent brain gene expression and copy number anomalies in autism suggest distinct pathological processes at young versus mature ages. PLoS Genet. 2012;8:e1002592. doi: 10.1371/journal.pgen.1002592. [DOI] [PMC free article] [PubMed] [Google Scholar]
Colombo J, Brez CC, Curtindale LM (2012) Infant perception and cognition. In: Lerner RM, Easterbrooks MA, Mistry J (eds) Handbook of psychology, vol 6: developmental psychology. Wiley, Hoboken, NJ, pp 61–89
Corbetta M, Shulman GL. Control of goal-directed and stimulus-driven attention in the brain. Nat Rev Neurosci. 2002;3:201–215. doi: 10.1038/nrn755. [DOI] [PubMed] [Google Scholar]
Courchesne E, Pierce K. Why the frontal cortex in autism might be talking only to itself: local over-connectivity but long-distance disconnection. Curr Opin Neurobiol. 2005;15:225–230. doi: 10.1016/j.conb.2005.03.001. [DOI] [PubMed] [Google Scholar]
Courchesne E, Redcay E, Morgan JT, Kennedy DP. Autism at the beginning: microstructural and growth abnormalities underlying the cognitive and behavioral phenotype of autism. Dev Psychopathol. 2005;17:577–597. doi: 10.1017/S0954579405050285. [DOI] [PubMed] [Google Scholar]
Courchesne E, Pierce K, Schumann CM, Redcay E, Buckwalter JA, Kennedy DP, Morgan JT. Mapping early brain development in autism. Neuron. 2007;56:399–413. doi: 10.1016/j.neuron.2007.10.016. [DOI] [PubMed] [Google Scholar]
Crespi B, Badcock C. Psychosis and autism as diametrical disorders of the social brain. Behav Brain Sci. 2008;31:241–320. doi: 10.1017/S0140525X08004214. [DOI] [PubMed] [Google Scholar]
Currenti SA. Understanding and determining the etiology of autism. Cell Mol Neurobiol. 2010;30:161–171. doi: 10.1007/s10571-009-9453-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
Damoiseaux JS, Rombouts SARB, Barkhof F, Scheltens P, Stam CJ, Smith SM, Beckmann CF. Consistent resting-state networks across healthy subjects. Proc Natl Acad Sci USA. 2006;103:13848–13853. doi: 10.1073/pnas.0601417103. [DOI] [PMC free article] [PubMed] [Google Scholar]
Dehaene S, Changeux JP. Neural mechanisms for access to consciousness. In: Gazzaniga MS, editor. The cognitive neurosciences. 3. Cambridge, MA: MIT Press; 2004. pp. 1145–1157. [Google Scholar]
Dehaene S, Changeux JP. Experimental and theoretical approaches to conscious processing. Neuron. 2011;70:200–227. doi: 10.1016/j.neuron.2011.03.018. [DOI] [PubMed] [Google Scholar]
Dehaene S, Naccache L. Towards a cognitive neuroscience of consciousness: basic evidence and a workspace framework. Cognition. 2001;79:1–37. doi: 10.1016/S0010-0277(00)00123-2. [DOI] [PubMed] [Google Scholar]
Dehaene S, Charles L, King J-R, Marti S. Toward a computational theory of conscious processing. Curr Opin Neurobiol. 2014;25:76–84. doi: 10.1016/j.conb.2013.12.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
Dehaene-Lambertz G, Spelke ES. The infancy of the human brain. Neuron. 2015;88:93–109. doi: 10.1016/j.neuron.2015.09.026. [DOI] [PubMed] [Google Scholar]
DeRamus TP, Black BS, Pennick MR, Kana RK. Enhanced parietal cortex activation during location detection in children with autism. Neurodev Disord. 2014;6:37. doi: 10.1186/1866-1955-6-37. [DOI] [PMC free article] [PubMed] [Google Scholar]
Di Martino A, Fair DA, Kelly C, Satterthwaite TD, Castellanos FX, Thomason ME, Craddock RC, Luna B, Leventhal BL, Zhuo X-N, Milham MP. Unraveling the miswired connectome: a developmental perspective. Neuron. 2014;83:1335–1353. doi: 10.1016/j.neuron.2014.08.050. [DOI] [PMC free article] [PubMed] [Google Scholar]
Dinstein I, Pierce K, Eyler L, Solso S, Malach R, Behrmann M, Courchesne E. Disrupted neural synchronization in toddlers with autism. Neuron. 2011;70:1218–1225. doi: 10.1016/j.neuron.2011.04.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
Doria V, Beckmann CF, Arichi T, Merchant N, Groppo M, Turkheimer FE, Counsell SJ, Murgasova M, Aljabar P, Nunes RG, Larkman DJ, Rees G, Edwards AD. Emergence of resting state networks in the preterm human brain. Proc Natl Acad Sci USA. 2010;107:20015–20020. doi: 10.1073/pnas.1007921107. [DOI] [PMC free article] [PubMed] [Google Scholar]
Elsabbagh M, Fernandes J, Webb SJ, Dawson G, Charman T, Johnson MH, The British Autism Study of Infant Siblings Team Disengagement of visual attention in infancy is associated with emerging autism in toddlerhood. Biol Psychol. 2013;74:189–194. doi: 10.1016/j.biopsych.2012.11.030. [DOI] [PMC free article] [PubMed] [Google Scholar]
Elton A, Di Martino A, Hazlett HC, Gao W. Neural connectivity evidence for a categorical-dimensional hybrid model of autism spectrum disorder. Biol Psychiatry. 2016;80:120–128. doi: 10.1016/j.biopsych.2015.10.020. [DOI] [PMC free article] [PubMed] [Google Scholar]
Fair DA, Cohen AL, Dosenbach NUF, Church JA, Miezin FM, Barch DM, Raichle MA, Petersen SE, Schlaggar BL. The maturing architecture of the brain’s default network. Proc Natl Acad Sci USA. 2008;105:4028–4032. doi: 10.1073/pnas.0800376105. [DOI] [PMC free article] [PubMed] [Google Scholar]
Fernandino L, Binder JR, Desai RH, Pendl SL, Humphries CJ, Gross WL, Conant LL, Seidenberg MS. Concept representation reflects multimodal abstraction: a framework for embodied semantics. Cortex. 2015 doi: 10.1093/cercor/bhv020. [DOI] [PMC free article] [PubMed] [Google Scholar]
Fields C. Trajectory recognition as the basis for object individuation: a functional model of object file instantiation and object-token encoding. Front Psychol. 2011;2:49. doi: 10.3389/fpsyg.2011.00049. [DOI] [PMC free article] [PubMed] [Google Scholar]
Fields C. Do autism spectrum disorders involve a generalized object categorization and identification dysfunction? Med Hypotheses. 2012;79:344–351. doi: 10.1016/j.mehy.2012.05.032. [DOI] [PubMed] [Google Scholar]
Fields C. The very same thing: extending the object token concept to incorporate causal constraints on individual identity. Adv Cogn Psychol. 2012;8:234–247. doi: 10.5709/acp-0119-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
Fields C. The principle of persistence, Leibniz’s law, and the computational task of object re-identification. Hum Dev. 2013;56:147–166. doi: 10.1159/000350206. [DOI] [Google Scholar]
Fields C. Motion, identity and the bias toward agency. Front Hum Neurosci. 2014;8:597. doi: 10.3389/fnhum.2014.00597. [DOI] [PMC free article] [PubMed] [Google Scholar]
Fornito A, Zalesky A, Bassett DS, Meunier D, Ellison-Wright I, Yücel M, Wood SJ, Shaw K, O’Connor J, Nertney D, Mowry BJ, Pantelis C, Bullmore ET. Genetic influences on cost-efficient organization of human cortical functional networks. J Neurosci. 2011;31:3261–3270. doi: 10.1523/JNEUROSCI.4858-10.2011. [DOI] [PMC free article] [PubMed] [Google Scholar]
Franklin S, Ramamurthy U, D’Mello SK, McCauley L, Negatu A, Silva LR, Datla V (2007) LIDA: a computational model of global workspace theory and developmental learning. In: Proceedings of the AAAI fall symposium on artificial intelligence and consciousness. AAAI, Menlo Park, CA, pp 61–66
Friston KJ. The free-energy principle: a unified brain theory? Nat Rev Neurosci. 2010;11:127–138. doi: 10.1038/nrn2787. [DOI] [PubMed] [Google Scholar]
Friston KJ. Functional and effective connectivity: a review. Brain Connect. 2011;1:13–36. doi: 10.1089/brain.2011.0008. [DOI] [PubMed] [Google Scholar]
Friston K, Lawson R, Frith C. On hyperpriors and hypopriors: comment on Pellicano and Burr. Trends Cogn Sci. 2012;17:1. doi: 10.1016/j.tics.2012.11.003. [DOI] [PubMed] [Google Scholar]
Gallivan JP, Culham JC. Neural coding within human brain areas involved in actions. Curr Opin Neurobiol. 2015;33:141–149. doi: 10.1016/j.conb.2015.03.012. [DOI] [PubMed] [Google Scholar]
Gao W, Gilmore JH, Shen D, Smith JK, Zhu H, Lin W. The synchronization within and interaction between the default and dorsal attention networks in early infancy. Cereb Cortex. 2013;23:594–603. doi: 10.1093/cercor/bhs043. [DOI] [PMC free article] [PubMed] [Google Scholar]
Gao W, Alcauter S, Smith JK, Gilmore JH, Lin W. Development of human brain cortical network architecture during infancy. Brain Struct Funct. 2015;220:1173–1186. doi: 10.1007/s00429-014-0710-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
Gardener H, Spiegelman D, Buka HL. Perinatal and neonatal risk factors for autism: a comprehensive meta-analysis. Pediatrics. 2011;128:344–355. doi: 10.1542/peds.2010-1036. [DOI] [PMC free article] [PubMed] [Google Scholar]
Geschwind DH, Levitt S. Autism spectrum disorders: developmental disconnection syndromes. Curr Opin Neurobiol. 2007;17:103–111. doi: 10.1016/j.conb.2007.01.009. [DOI] [PubMed] [Google Scholar]
Geschwind DH. Autism: many genes, common pathways? Cell. 2008;135:391–395. doi: 10.1016/j.cell.2008.10.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
Geschwind DH. Advances in autism. Annu Rev Med. 2009;60:367–380. doi: 10.1146/annurev.med.60.053107.121225. [DOI] [PMC free article] [PubMed] [Google Scholar]
Geschwind DH, Flint J. Genetics and genomics of psychiatric disease. Science. 2015;349:1489–1494. doi: 10.1126/science.aaa8954. [DOI] [PMC free article] [PubMed] [Google Scholar]
Gibson JJ. The ecological approach to visual perception. Boston: Houghton Mifflin; 1979. [Google Scholar]
Glazebrook JF, Wallace R. Small worlds and red queens in the global workspace: an information-theoretic approach. Cogn Syst Res. 2009;10:333–365. doi: 10.1016/j.cogsys.2009.01.002. [DOI] [Google Scholar]
Glazebrook JF, Wallace R. Pathologies in functional connectivity, feedback control and robustness: a global workspace perspective on autism spectrum disorders. Cogn Proc. 2015;16:1–16. doi: 10.1007/s10339-014-0636-y. [DOI] [PubMed] [Google Scholar]
Gómez CM, Flores A. A neurophysiological evaluation of a cognitive cycle in humans. Neurosci Biobehav Rev. 2011;35:452–461. doi: 10.1016/j.neubiorev.2010.05.005. [DOI] [PubMed] [Google Scholar]
Gomot M, Wicker B. A challenging, unpredictable world for people with autism spectrum disorder. Int J Psychophysiol. 2012;83:240–247. doi: 10.1016/j.ijpsycho.2011.09.017. [DOI] [PubMed] [Google Scholar]
Gopnik A. Scientific thinking in young children: theoretical advances, empirical research, and policy implications. Science. 2012;337:1623–1627. doi: 10.1126/science.1223416. [DOI] [PubMed] [Google Scholar]
Gopnik A, Wellman HM. Reconstructing constructivism: causal models, Bayesian learning mechanisms and the theory theory. Psychol Bull. 2012;138:1085–1108. doi: 10.1037/a0028044. [DOI] [PMC free article] [PubMed] [Google Scholar]
Gottlieb J, Oudeyer P-Y, Lopes M, Baranes A. Information-seeking, curiosity, and attention: computational and neural mechanisms. Trends Cogn Sci. 2013;17:585–593. doi: 10.1016/j.tics.2013.09.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
Graziano MSA. Speculations of the evolution of awareness. J Cogn Neurosci. 2014;26:1300–1304. doi: 10.1162/jocn_a_00623. [DOI] [PubMed] [Google Scholar]
Grossberg S, Seidman D. Neural dynamics of autistic behaviors: cognitive, emotional, and timing substrates. Psychol Rev. 2006;113:483–525. doi: 10.1037/0033-295X.113.3.483. [DOI] [PubMed] [Google Scholar]
Hahamy A, Behrmann M, Malach R. The idiosyncratic brain: distortion of spontaneous connectivity patterns in autism spectrum disorder. Nat Neurosci. 2015;18:302–309. doi: 10.1038/nn.3919. [DOI] [PubMed] [Google Scholar]
Happé F, Frith U. The weak coherence account: detail-focused cognitive style in autism spectrum disorders. J Autism Dev Disord. 2006;36:5–25. doi: 10.1007/s10803-005-0039-0. [DOI] [PubMed] [Google Scholar]
Harms MB, Martin A, Wallace GL. Facial emotion recognition in autism spectrum disorders: a review of behavioral and neuroimaging studies. Neuropsychol Rev. 2010;20:290–322. doi: 10.1007/s11065-010-9138-6. [DOI] [PubMed] [Google Scholar]
Hellendoorn A, Wijnroks L, Leseman PPM. Unraveling the nature of autism: finding order amid change. Front Psychol. 2015;6:359. doi: 10.3389/fpsyg.2015.00359. [DOI] [PMC free article] [PubMed] [Google Scholar]
Hoehl S. The development of category specificity in infancy—what can we learn from electrophysiology? Neuropsychologia. 2016;83:114–122. doi: 10.1016/j.neuropsychologia.2015.08.021. [DOI] [PubMed] [Google Scholar]
Hoelh S, Peykarjou S. The early development of face processing—what makes faces special? Neurosci Bull. 2012;28:765–788. doi: 10.1007/s12264-012-1280-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
Hoffman D, Prakash C. Objects of consciousness. Front Psychol. 2014;5:577. doi: 10.3389/fpsyg.2014.00577. [DOI] [PMC free article] [PubMed] [Google Scholar]
Huang H, Shu N, Mishra V, Jeon T, Chalak L, Wang ZJ, Rollins N, Gong G, Cheng H, Peng Y, Dong Q, He Y. Development of human brain structural networks through infancy and childhood. Cereb Cortex. 2015;25:1389–1404. doi: 10.1093/cercor/bht335. [DOI] [PMC free article] [PubMed] [Google Scholar]
Iacoboni M, Dapretto M. The mirror neuron system and the consequences of its dysfunction. Nat Rev Neurosci. 2006;7:942–951. doi: 10.1038/nrn2024. [DOI] [PubMed] [Google Scholar]
Ishibashi R, Pobric G, Saito S, Lambon Ralph MA. The neural network for tool-related cognition: an activation likelihood estimation meta-analysis of 70 neuroimaging contrasts. Cogn Neuropsychol. 2016;33:241–256. doi: 10.1080/02643294.2016.1188798. [DOI] [PMC free article] [PubMed] [Google Scholar]
Jarman N, Trengove C, Steur E, Tyukin I, van Leeuwen C. Spatially constrained adaptive rewiring in cortical networks creates spatially modular small world architectures. Cogn Neurodyn. 2014;8:479–497. doi: 10.1007/s11571-014-9288-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
Jeste SS. The neurology of autism spectrum disorders. Curr Opin Neurol. 2011;24:132–139. doi: 10.1097/WCO.0b013e3283446450. [DOI] [PMC free article] [PubMed] [Google Scholar]
Johnson MH. Interactive specialization: a domain-general framework for human functional brain development? Dev Cogn Neurosci. 2011;1:7–21. doi: 10.1016/j.dcn.2010.07.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
Just M, Keller T, Malave V, Kana R, Varma S. Autism as a neural disorder: a theory of frontal-posterior underconnectivity. Neurosci Biobehav Rev. 2012;36:1292–1313. doi: 10.1016/j.neubiorev.2012.02.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
Kaiser MD, Pelphrey KA. Disrupted action perception in autism: behavioral evidence, neuroendophenotypes, and diagnostic utility. Dev Cogn Neurosci. 2012;2:25–35. doi: 10.1016/j.dcn.2011.05.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
Keehn B, Wagner JB, Tager-Flusberg H, Nelson CA. Functional connectivity in the first year of life in infants at-risk for autism: a preliminary near-infrared spectroscopy study. Front Human Neurosci. 2013;7:444. doi: 10.3389/fnhum.2013.00444. [DOI] [PMC free article] [PubMed] [Google Scholar]
Kennedy DP, Semendeferi K, Courchesne E. No reduction of spindle neuron number in frontoinsular cortex in autism. Brain Cogn. 2007;64:124–129. doi: 10.1016/j.bandc.2007.01.007. [DOI] [PubMed] [Google Scholar]
Kidd C, Piantadosi ST, Aslin RN. The Goldilocks effect: human infants allocate attention to visual sequences that are neither too simple nor too complex. PLoS ONE. 2012;7:e36399. doi: 10.1371/journal.pone.0036399. [DOI] [PMC free article] [PubMed] [Google Scholar]
Kiefer M, Pulvermüller F. Conceptual representations in mind and brain: theoretical developments, current evidence and future directions. Cortex. 2012;48:805–825. doi: 10.1016/j.cortex.2011.04.006. [DOI] [PubMed] [Google Scholar]
Kita EM, Scott EK, Goodhill GJ. The influence of activity on axon pathfinding in the optic tectum. Dev Neurobiol. 2015;75:608–620. doi: 10.1002/dneu.22262. [DOI] [PubMed] [Google Scholar]
Knill DC, Pouget A. The Bayesian brain: the role of uncertainty in neural coding and computation. Trends Neurosci. 2004;27:712–719. doi: 10.1016/j.tins.2004.10.007. [DOI] [PubMed] [Google Scholar]
Latora V, Marchiori M. Efficient behavior of small-world networks. Phys Rev Lett. 2001;87:198701. doi: 10.1103/PhysRevLett.87.198701. [DOI] [PubMed] [Google Scholar]
Lawson RP, Rees G, Friston KJ. An aberrant precision account of autism. Front Hum Neurosci. 2014;8:302. doi: 10.3389/fnhum.2014.00302. [DOI] [PMC free article] [PubMed] [Google Scholar]
Leonard H, Dixon G, Whitehouse AJO, et al. Unpacking the complex nature of the autism epidemic. Res Autism Spectrum Disord. 2010;4:548–554. doi: 10.1016/j.rasd.2010.01.003. [DOI] [Google Scholar]
Lewis JW. Cortical networks related to human use of tools. Neuroscientist. 2006;12:211–231. doi: 10.1177/1073858406288327. [DOI] [PubMed] [Google Scholar]
Lewis JD, Evans AC, Pruett JR, Botterton K, Zwaigenbaum L, Estes A, Gerig G, Collins L, Kostopoulos P, McKinstry R, Dages S, Paterson S, Schultz RT, Styner M, Hazlett H, Piven J, (IBIS Network) Network inefficiencies in autism spectrum disorder at 24 months. Transl Psychiatry. 2014;4:e388. doi: 10.1038/tp.2014.24. [DOI] [PMC free article] [PubMed] [Google Scholar]
Lombardo MV, Pierce K, Eyler LT, Barnes CC, Ahrens-Barbeau C, Solso S, Campbell K, Courchesne E. Different functional neural substrates for good and poor language outcome in autism. Neuron. 2015;86:567–577. doi: 10.1016/j.neuron.2015.03.023. [DOI] [PMC free article] [PubMed] [Google Scholar]
Luo Y. Three-month-old infants attribute goals to a non-human agent. Dev Sci. 2011;14:453–460. doi: 10.1111/j.1467-7687.2010.00995.x. [DOI] [PubMed] [Google Scholar]
Luo Y, Baillargeon R. Toward a mentalistic account of early psychological reasoning. Curr Dir Psychol Sci. 2010;19:301–307. doi: 10.1177/0963721410386679. [DOI] [PMC free article] [PubMed] [Google Scholar]
Markram H, Rinaldi T, Markram K. The intense world syndrome—an alternative hypothesis for autism. Front Neurosci. 2007;1:77–96. doi: 10.3389/neuro.01.1.1.006.2007. [DOI] [PMC free article] [PubMed] [Google Scholar]
Matson JL, Kozlowski AM. The increasing prevalence of autism spectrum disorders. Res Autism Spectrum Disord. 2011;5:418–425. doi: 10.1016/j.rasd.2010.06.004. [DOI] [Google Scholar]
Maximo JO, Cadena EJ, Kana RK. The implications of brain connectivity in the neuropsychology of autism. Neuropsychol Rev. 2014;24:16–31. doi: 10.1007/s11065-014-9250-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
McCall R, Franklin S (2013) Cortical learning algorithms with predictive coding for a systems-level cognitive architecture. In: Proceedings of the second annual conference on advances in cognitive systems. Cognitive Systems Foundation, pp 149–166
McFadden K, Minshew NJ. Evidence for dysregulation of axonal growth and guidance in the etiology of ASD. Front Hum Neurosci. 2013;7:671. doi: 10.3389/fnhum.2013.00671. [DOI] [PMC free article] [PubMed] [Google Scholar]
Miles JH. Autism subgroups from a medical genetics perspective. In: Amaral DG, Dawson G, Geschwind DH, editors. Autism spectrum disorders. New York: Oxford University Press; 2011. pp. 705–721. [Google Scholar]
Minshew NJ, Williams DL. The new neurobiology of autism. Arch Neurol. 2007;64:945–950. doi: 10.1001/archneur.64.7.945. [DOI] [PMC free article] [PubMed] [Google Scholar]
Müller R-A, Shih P, Keehn B, Deyoe JR, Leyden KN, Shukla DK. Underconnected, but how? A survey of functional connectivity MRI studies in autism spectrum disorders. Cereb Cortex. 2011;21:2233–2243. doi: 10.1093/cercor/bhq296. [DOI] [PMC free article] [PubMed] [Google Scholar]
Munz M, Gobert D, Schohl A, Poquérusse J, Podgorski K, Spratt P, Ruthazer ES. Rapid Hebbian axonal remodeling mediated by visual stimulation. Science. 2014;344:904–909. doi: 10.1126/science.1251593. [DOI] [PubMed] [Google Scholar]
Needham A, Dueker G, Lockhead G. Infants’ formation and use of categories to segregate objects. Cognition. 2005;94:215–240. doi: 10.1016/j.cognition.2004.02.002. [DOI] [PubMed] [Google Scholar]
Newman MEJ. The structure and function of complex networks. SIAM Rev. 2003;45:167–256. doi: 10.1137/S003614450342480. [DOI] [Google Scholar]
Oberman LM, Ramachandran VS. The simulating social mind: the role of the mirror neuron system and simulation in the social and communicative deficits of autism spectrum disorders. Psychol Bull. 2007;133:310–327. doi: 10.1037/0033-2909.133.2.310. [DOI] [PubMed] [Google Scholar]
Osiurak F, Jarry C, Lesourd M, Baumard J, Le Gall D. Mechanical problem-solving strategies in left-brain damaged patients and apraxia of tool use. Neuropsychology. 2013;51:1964–1972. doi: 10.1016/j.neuropsychologia.2013.06.017. [DOI] [PubMed] [Google Scholar]
Oudeyer P-Y, Baranes A, Kaplan F. Intrinsically motivated learning of real world sensorimotor skills with developmental constraints. In: Baldassarre G, Mirolli M, editors. Intrinsically motivated learning in natural and artificial systems. Berlin: Springer; 2013. pp. 303–365. [Google Scholar]
Parikshak NN, Luo R, Zhang A, Won H, Lowe JK, Chandran V, Horvath S, Geschwind DH. Integrative functional genomic analyses implicate specific molecular pathways and circuits in autism. Cell. 2013;155:1008–1021. doi: 10.1016/j.cell.2013.10.031. [DOI] [PMC free article] [PubMed] [Google Scholar]
Park H-J, Friston K. Structural and functional brain networks: from connections to cognition. Science. 2013;342:1238411-1–1238411-8. doi: 10.1126/science.1238411. [DOI] [PubMed] [Google Scholar]
Peca J, Feng G. Cellular and synaptic network defects in autism. Curr Opin Neurol. 2012;22:866–872. doi: 10.1016/j.conb.2012.02.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
Pellicano E, Burr D. When the world becomes `too real’: a Bayesian explanation of autistic perception. Trends Cogn Sci. 2012;16:504–510. doi: 10.1016/j.tics.2012.08.009. [DOI] [PubMed] [Google Scholar]
Pelphrey KA, Shultz S, Hudac CM, Vander Wyk BC. Constraining heterogeneity: the social brain and its development in autism spectrum disorder. J Child Psychol Psychiatry. 2011;52:631–644. doi: 10.1111/j.1469-7610.2010.02349.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
Peters JM, Taquet M, Vega C, et al. Brain functional networks in syndromic and non-syndromic autism: a graph theoretical study of EEG connectivity. BMC Med. 2013;11:54. doi: 10.1186/1741-7015-11-54. [DOI] [PMC free article] [PubMed] [Google Scholar]
Picci G, Gotts SJ, Scherf KS. A theoretical rut: revisiting and critically evaluating the generalized under/over-connectivity hypothesis of autism. Dev Sci. 2016;19:524–549. doi: 10.1111/desc.12467. [DOI] [PubMed] [Google Scholar]
Pinto D, Delaby E, Merico D, et al. Convergence of genes and cellular pathways dysregulated in autism spectrum disorders. Am J Hum Genet. 2014;94:677–694. doi: 10.1016/j.ajhg.2014.03.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
Rajendran G, Mitchell P. Cognitive theories of autism. Dev Rev. 2007;27:224–260. doi: 10.1016/j.dr.2007.02.001. [DOI] [Google Scholar]
Rakison DH, Yermoleva Y. Infant categorization. Wiley Interdiscip Rev Cogn Sci. 2010;1:894–905. doi: 10.1002/wcs.81. [DOI] [PubMed] [Google Scholar]
Righi G, Tierney AL, Tager-Flusberg H, Nelson CA. Functional connectivity in the first year of life in infants at risk for autism spectrum disorder: an EEG study. PLoS ONE. 2014;9:e105176. doi: 10.1371/journal.pone.0105176. [DOI] [PMC free article] [PubMed] [Google Scholar]
Rippon G, Brock J, Brown C, Boucher J. Disordered connectivity in the autistic brain: challenges for the ‘new psychophysiology’. Int J Psychophysiol. 2007;63:164–172. doi: 10.1016/j.ijpsycho.2006.03.012. [DOI] [PubMed] [Google Scholar]
Robertson CE, Thomas C, Kravitz DJ, Wallace GL, Baron-Cohen S, Martin A, Baker CI. Global motion perception deficits in autism are reflected as early as primary visual cortex. Brain. 2014;137:2588–2599. doi: 10.1093/brain/awu189. [DOI] [PMC free article] [PubMed] [Google Scholar]
Rochat P. Primordial sense of embodied self-unity. In: Slaughter V, Brownell CA, editors. Early development of body representations. Cambridge: Cambridge University Press; 2012. pp. 3–18. [Google Scholar]
Rolls ET. Limbic systems for emotion and for memory, but no single limbic system. Cortex. 2015;62:119–157. doi: 10.1016/j.cortex.2013.12.005. [DOI] [PubMed] [Google Scholar]
Rubinov M, Sporns O. Complex network measures of brain connectivity: uses and interpretations. NeuroImage. 2010;52:1059–1069. doi: 10.1016/j.neuroimage.2009.10.003. [DOI] [PubMed] [Google Scholar]
Rugg MD, Vilberg KL. Brain networks underlying episodic memory retrieval. Curr Opin Neurobiol. 2013;23:255–260. doi: 10.1016/j.conb.2012.11.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
Santos M, Uppal N, Butti C, Wicinski B, Schmeidler J, Giannakopoulos P, Heinsen H, Schmitz C, Hof PR. von Economo neurons in autism: a stereologic study of the frontoinsular cortex in children. Brain Res. 2011;1380:206–217. doi: 10.1016/j.brainres.2010.08.067. [DOI] [PubMed] [Google Scholar]
Schlesinger M, Amso D, Johnson SP, Hantehzadeh N, Gupta L (2012) Using the iCub simulator to study perceptual development: a case study. In: Proceedings of the 2012 IEEE international conference on development and learning and epigenetic robotics (ICDL). IEEE, pp 1–6. doi:10.1109/DevLrn.2012.6400866
Shanahan M. The brain’s connective core and its role in animal cognition. Philos Trans R Soc B. 2012;367:2704–2714. doi: 10.1098/rstb.2012.0128. [DOI] [PMC free article] [PubMed] [Google Scholar]
Shih P, Shen M, Öttl B, Keehn B, Gaffrey MS, Müller R-A. Atypical network connectivity for imitation in autism spectrum disorder. Neuropsychologia. 2010;48:2931–2939. doi: 10.1016/j.neuropsychologia.2010.05.035. [DOI] [PMC free article] [PubMed] [Google Scholar]
Shipp S, Adams RA, Friston KJ. Reflections on agranular architecture: predictive coding in the motor cortex. Trends Neurosci. 2013;36:706–716. doi: 10.1016/j.tins.2013.09.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
Simion F, Di Giorgio E, Leo I, Bardi L. The processing of social stimuli in early infancy: from faces to biological motion. Prog Brain Res. 2011;189:173–193. doi: 10.1016/B978-0-444-53884-0.00024-5. [DOI] [PubMed] [Google Scholar]
Simms ML, Kemper TL, Timbie CM, Bauman ML, Blatt GJ. The anterior cingulate cortex in autism: heterogeneity of qualitative and quantitative cytoarchitectonic features suggests possible subgroups. Acta Neuropathol. 2009;118:673–684. doi: 10.1007/s00401-009-0568-2. [DOI] [PubMed] [Google Scholar]
Sporns O. Network attributes for segregation and integration in the human brain. Curr Opin Neurobiol. 2013;23:162–171. doi: 10.1016/j.conb.2012.11.015. [DOI] [PubMed] [Google Scholar]
Sporns O, Honey CJ. Small worlds inside big brains. Proc Natl Acad Sci USA. 2006;111:15220–15225. doi: 10.1073/pnas.0609523103. [DOI] [PMC free article] [PubMed] [Google Scholar]
Sporns O, Zwi JD. The small world of the cerebral cortex. Neuroinformatics. 2004;2:145–162. doi: 10.1385/NI:2:2:145. [DOI] [PubMed] [Google Scholar]
Sporns O, Tononi G, Edelman GM. Theoretical neuroanatomy and the connectivity of the cerebral cortex. Behav Brain Res. 2002;135:69–74. doi: 10.1016/S0166-4328(02)00157-2. [DOI] [PubMed] [Google Scholar]
Stahl AE, Feigenson L. Observing the unexpected enhances infants’ learning and exploration. Science. 2015;348:91–94. doi: 10.1126/science.aaa3799. [DOI] [PMC free article] [PubMed] [Google Scholar]
Sunderland A, Wilkins L, Dineen R, Dawson SE. Tool-use and the left hemisphere: what is lost in ideomotor apraxia? Brain Cogn. 2013;81:183–192. doi: 10.1016/j.bandc.2012.10.008. [DOI] [PubMed] [Google Scholar]
Tanaka JW, Sung A. The “eye avoidance” hypothesis of autism face processing. J Autism Dev Disord. 2016;46:1538–1552. doi: 10.1007/s10803-013-1976-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
Tononi G, Koch C. Consciousness: here, there and everywhere? Philos Trans R Soc B. 2015;370:20140167. doi: 10.1098/rstb.2014.0167. [DOI] [PMC free article] [PubMed] [Google Scholar]
Toulmin H, Beckmann CF, O’Muircheartaigh J, Ball G, Nongena P, Makropoulos A, Ederies A, Counsell SJ, Kennea N, Arichi T, Tusor N, Rutherford MA, Azzopardi D, Gonzalez-Cinca N, Hajnal JV, Edwards AD. Specialization and integration of functional thalamocortical connectivity in the human infant. Proc Natl Acad Sci USA. 2015;112:6485–6490. doi: 10.1073/pnas.1422638112. [DOI] [PMC free article] [PubMed] [Google Scholar]
Tyszka JM, Kennedy DP, Paul LK, Adolphs R. Largely typical patterns of resting-state functional connectivity in high-functioning adults with autism. Cereb Cortex. 2014;24:1894–1905. doi: 10.1093/cercor/bht040. [DOI] [PMC free article] [PubMed] [Google Scholar]
Uddin LQ. Salience processing and insular cortical function and dysfunction. Nat Rev Neurosci. 2015;16:55–61. doi: 10.1038/nrn3857. [DOI] [PubMed] [Google Scholar]
Uddin LQ, Supekar K, Lynch CJ, Khouzam A, Phillips J, Feinsten C, Ryall S, Menon V. Salience network-based classification and prediction of symptom severity in children with autism. JAMA Psychiatry. 2013;70:869–879. doi: 10.1001/jamapsychiatry.2013.104. [DOI] [PMC free article] [PubMed] [Google Scholar]
Uddin LQ, Supekar K, Menon V. Reconceptualizing functional brain connectivity in autism from a developmental perspective. Front Hum Neurosci. 2013;7:458. doi: 10.3389/fnhum.2013.00458. [DOI] [PMC free article] [PubMed] [Google Scholar]
Vaccarino FM, Grigorenko EL, Smith KM, Stevens H. Regulation of cerebral cortical size and neuron number by fibroblast growth factors: implications for autism. J. Autism Dev Disord. 2009;39:511–520. doi: 10.1007/s10803-008-0653-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
van Boxtel JJA, Lu H. A predictive coding perspective on autism spectrum disorders. Front Psychol. 2013;4:19. doi: 10.3389/fpsyg.2013.00019. [DOI] [PMC free article] [PubMed] [Google Scholar]
Van de Cruys S, Evers K, Van der Hallen R, Van Eylen L, Boets B, de-Wit L, Wagemans J. Precise minds in uncertain worlds: predictive coding in autism. Psychol Rev. 2014;121:649–675. doi: 10.1037/a0037665. [DOI] [PubMed] [Google Scholar]
van den Heuvel MP, Sporns O. Rich-club organization of the human connectome. J Neurosci. 2011;31:15775–15786. doi: 10.1523/JNEUROSCI.3539-11.2011. [DOI] [PMC free article] [PubMed] [Google Scholar]
van Leeuwen C. What makes you think you are conscious? An agnosticist manifesto. Front Hum Neurosci. 2015;9:170. doi: 10.3389/fnhum.2015.00170. [DOI] [PMC free article] [PubMed] [Google Scholar]
Vértes PE, Bullmore ET. Annual research review: growth connectomics—the organization and reorganization of brain networks during normal and abnormal development. J Child Psychol Psychiatry. 2015;56:299–320. doi: 10.1111/jcpp.12365. [DOI] [PMC free article] [PubMed] [Google Scholar]
Vingerhoets G. Contribution of the posterior parietal cortex in reaching, grasping, and using objects and tools. Front Psychol. 2014;5:151. doi: 10.3389/fpsyg.2014.00151. [DOI] [PMC free article] [PubMed] [Google Scholar]
Vissers ME, Cohen MX, Geurts HM. Brain connectivity and high functioning autism: a promising path of research that needs refined models, methodological convergence, and strong behavioral links. Neurosci Biobehav Rev. 2012;36:604–625. doi: 10.1016/j.neubiorev.2011.09.003. [DOI] [PubMed] [Google Scholar]
Volkmar F, Chawarska K, Klin A. Autism in infancy and early childhood. Annu Rev Psychol. 2005;56:315–336. doi: 10.1146/annurev.psych.56.091103.070159. [DOI] [PubMed] [Google Scholar]
von Hofsten C. Action in development. Dev Sci. 2007;10:54–60. doi: 10.1111/j.1467-7687.2007.00564.x. [DOI] [PubMed] [Google Scholar]
Vossel S, Geng JJ, Fink GR. Dorsal and ventral attention systems: distinct neural circuits but collaborative roles. Neurosci. 2014;20:150–159. doi: 10.1177/1073858413494269. [DOI] [PMC free article] [PubMed] [Google Scholar]
Wallace R. Consciousness: a mathematical treatment of the global neuronal workspace model. New York: Springer; 2005. [Google Scholar]
Wass S. Distortions and disconnections: disrupted brain connectivity in autism. Brain Cogn. 2011;75:18–28. doi: 10.1016/j.bandc.2010.10.005. [DOI] [PubMed] [Google Scholar]
Watts DJ, Strogatz SH. Collective dynamics of ‘small-world’ networks. Nature. 1998;393:440–442. doi: 10.1038/30918. [DOI] [PubMed] [Google Scholar]
Weigelt S, Koldewyn K, Kanwisher N. Face identity recognition in autism spectrum disorders: a review of behavioral studies. Neurosci Biobehav Rev. 2012;36:1060–1084. doi: 10.1016/j.neubiorev.2011.12.008. [DOI] [PubMed] [Google Scholar]
Westermann G, Mareschal D. From perceptual to language-mediated categorization. Philos Trans R Soc B. 2012;369:20120391. doi: 10.1098/rstb.2012.0391. [DOI] [PMC free article] [PubMed] [Google Scholar]
Williams JH, Whiten A, Suddendorf T, Perrett DI. Imitation, mirror neurons and autism. Neurosci Biobehav Rev. 2001;25:287–295. doi: 10.1016/S0149-7634(01)00014-8. [DOI] [PubMed] [Google Scholar]
Willsey AJ, Sanders SJ, Li M, et al. Coexpression networks implicate human midfetal deep cortical projection neurons in the pathogenesis of autism. Cell. 2013;155:997–1007. doi: 10.1016/j.cell.2013.10.020. [DOI] [PMC free article] [PubMed] [Google Scholar]
Yap P-T, Fan Y, Chen Y, Gilmore JH, Lin W, Shen D. Development trends of white matter connectivity in the first years of life. PLoS ONE. 2011;6:e24678. doi: 10.1371/journal.pone.0024678. [DOI] [PMC free article] [PubMed] [Google Scholar]
Yeo BTT, Krienen FM, Sepulcre J, Sabuncu MR, Lashkari D, Hollinshead M, Roffman JL, Smoller JW, Zöllei L, Polimeni JR, Fischl B, Liu H, Buckner RL. The organization of the human cerebral cortex estimated by intrinsic functional connectivity. J Neurophysiol. 2011;106:1125–1165. doi: 10.1152/jn.00338.2011. [DOI] [PMC free article] [PubMed] [Google Scholar]
Zimmer HD, Ecker UKD. Remembering perceptual features unequally bound in object and episodic tokens: neural mechanisms and their electrophysiological correlates. Neurosci Biobehav Rev. 2010;34:1066–1079. doi: 10.1016/j.neubiorev.2010.01.014. [DOI] [PubMed] [Google Scholar]
Zmigrod S, de Sonneville LMJ, Colzato LS, Swaab H, Hommel B. Cognitive control of feature bindings: evidence from children with autistic spectrum disorder. Psychol Res. 2013;77:147–154. doi: 10.1007/s00426-011-0399-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
Zwaigenbaum L, Young GS, Stone WL, Dobkins K, Ozonoff S, Brian J, Bryson SE, Carver LJ, Hutman T, Iverson JM, Landa RJ, Messinger D. Early head growth in infants at risk of autism: a baby siblings research consortium study. J Am Acad Child Adolesc Psychiatry. 2014;53:1053–1062. doi: 10.1016/j.jaac.2014.07.007. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR1] Abbott AE, Nair A, Keown CL, Datko M, Jahedi A, Fishman I, Müller R-A. Patterns of atypical functional connectivity and behavioral links in autism differ between default, salience, and executive networks. Cereb Cortex. 2015;26:4034–4045. doi: 10.1093/cercor/bhv191. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR2] Achard S, Salvador R, Whitcher B, Suckling J, Bullmore E. A resilient, low-frequency, small-world, human brain functional network with highly connected association cortical hubs. J Neurosci. 2006;26:63–72. doi: 10.1523/JNEUROSCI.3874-05.2006. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR3] Adams RA, Friston KJ, Bastos AM. Active inference, predictive coding and cortical architecture. In: Casanova MF, Opris I, editors. Recent advances in the modular organization of the cortex. Berlin: Springer; 2015. pp. 97–121. [Google Scholar]

[CR4] Alcauter S, Lin W, Smith JK, Short SJ, Goldman BD, Reznick JS, Gilmore JH, Gao W. Development of thalamocortical connectivity during infancy and its cognitive correlations. J Neurosci. 2014;34:9067–9075. doi: 10.1523/JNEUROSCI.0796-14.2014. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR5] Allman JM, Tetreault NA, Hakeem AY, Manaye KF, Semendeferi K, Erwin JM, Park S, Goubert V, Hof PR. The von Economo neurons in fronto-insular and anterior cingulate cortex. Ann N Y Acad Sci. 2011;1225:59–71. doi: 10.1111/j.1749-6632.2011.06011.x. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR6] American Psychiatric Association . Diagnostic and statistical manual of mental disorders (DSM-V) Arlington, VA: American Psychiatric Association; 2013. [Google Scholar]

[CR7] Andersen SL. Trajectories of brain development: point of vulnerability or window of opportunity? Neurosci Biobehav Rev. 2003;27:3–18. doi: 10.1016/S0149-7634(03)00005-8. [DOI] [PubMed] [Google Scholar]

[CR8] Andrews-Hanna JR, Smallwood J, Spreng RN. The default network and self-generated thought: component processes, dynamic control, and clinical relevance. Ann N Y Acad Sci. 2014;1316:29–52. doi: 10.1111/nyas.12360. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR9] Aslin RN. What’s in a look? Dev Sci. 2007;10:48–53. doi: 10.1111/j.1467-7687.2007.00563.x. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR10] Aslin RN. Infant learning: historical, conceptual, and methodological challenges. Infancy. 2014;19:2–27. doi: 10.1111/infa.12036. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR11] Baars B. A cognitive theory of consciousness. New York: Cambridge University Press; 1998. [Google Scholar]

[CR12] Baars BJ. Global workspace theory of consciousness: toward a cognitive neuroscience of human experience. Prog Brain Res. 2005;150:45–53. doi: 10.1016/S0079-6123(05)50004-9. [DOI] [PubMed] [Google Scholar]

[CR13] Baars BJ, Franklin S. How conscious experience and working memory interact. Trends Cogn Sci. 2003;7:166–172. doi: 10.1016/S1364-6613(03)00056-1. [DOI] [PubMed] [Google Scholar]

[CR14] Baars BJ, Franklin S, Ramsoy TZ. Global workspace dynamics: cortical “binding and propagation” enables conscious contents. Front Psychol. 2013;4:200. doi: 10.3389/fpsyg.2013.00200. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR15] Baillargeon R, Li J, Gertner Y, Wu D. How do infants reason about physical events? In: Goswami U, editor. The Wiley-Blackwell handbook of child cognitive development. 2. Oxford: Blackwell; 2011. pp. 11–48. [Google Scholar]

[CR16] Ball G, Alijabar P, Zebari S, Tusor N, Arichi T, Merchant N, Robinson EC, Ogundipe E, Ruekert D, Edwards AD, Counsell SJ. Rich-club organization of the newborn human brain. Proc Natl Acad Sci USA. 2014;111:7456–7461. doi: 10.1073/pnas.1324118111. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR17] Barabasi AL, Albert R. Emergence of scaling in random networks. Science. 1999;286:509–512. doi: 10.1126/science.286.5439.509. [DOI] [PubMed] [Google Scholar]

[CR18] Baron-Cohen S. The extreme male brain theory of autism. Trends Cogn Sci. 2002;2:248–254. doi: 10.1016/S1364-6613(02)01904-6. [DOI] [PubMed] [Google Scholar]

[CR19] Bassett DS, Bullmore E. Small world brain networks. Neurosci. 2006;12:512–523. doi: 10.1177/1073858406293182. [DOI] [PubMed] [Google Scholar]

[CR20] Bastos AM, Usrey WM, Adams RA, Mangun GR, Fries P, Friston KJ. Canonical microcircuits for predictive coding. Neuron. 2012;76:695–711. doi: 10.1016/j.neuron.2012.10.038. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR21] Baum SH, Stevenson RA, Wallace MT. Behavioral, perceptual, and neural alterations in sensory and multisensory function in autism spectrum disorder. Prog Neurobiol. 2015;134:140–160. doi: 10.1016/j.pneurobio.2015.09.007. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR22] Belmonte MK, Baron-Cohen S (2004) Small-world network properties and the emergence of social cognition: evidence from functional studies of autism. In: Triesch J, Jebara T (eds) Proceedings of the 2004 international conference on development and learning. UCSD Institute for Neural Computation, La Jolla, CA, p 268

[CR23] Betancur C. Etiological heterogeneity in autism spectrum disorders: more than 100 genetic and genomic disorders and still counting. Brain Res. 2011;1380:42–77. doi: 10.1016/j.brainres.2010.11.078. [DOI] [PubMed] [Google Scholar]

[CR24] Block N, Carmel D, Fkeming SM, Kentridge RW, Koch C, Lamme VAF, Lau H, Rosenthal D. Consciousness science: real progress and lingering misconceptions. Trends Cogn Sci. 2014;18:556–557. doi: 10.1016/j.tics.2014.09.004. [DOI] [PubMed] [Google Scholar]

[CR25] Brandi M-L, Wohlschläger A, Sorg C, Hermsdörfer J. The neural correlates of planning and executing actual tool use. J Neurosci. 2014;34:13183–13194. doi: 10.1523/JNEUROSCI.0597-14.2014. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR26] Bremner JG, Johnson SP, Slater A, Mason U, Foster K, Cheshire A, Spring J. Conditions for young infants’ perception of object trajectories. Child Dev. 2005;76:1029–1043. doi: 10.1111/j.1467-8624.2005.00895.x. [DOI] [PubMed] [Google Scholar]

[CR27] Butti C, Santos M, Uppal N, Hof PR. Von Economo neurons: clinical and evolutionary perspectives. Cortex. 2013;49:312–326. doi: 10.1016/j.cortex.2011.10.004. [DOI] [PubMed] [Google Scholar]

[CR28] Buxbaum LJ, Shapiro AD, Coslett HB. Critical brain regions for tool-related and imitative actions: a componential analysis. Brain. 2014;137:1971–1985. doi: 10.1093/brain/awu111. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR29] Cangelosi A, Schlesinger M. Developmental robotics: from babies to robots. Cambridge, MA: MIT Press; 2015. [Google Scholar]

[CR30] Casey BJ, Tottenham N, Liston C, Durstan S. Imaging the developing brain: what have we learned about cognitive development? Trends Cogn Sci. 2005;9:104–110. doi: 10.1016/j.tics.2005.01.011. [DOI] [PubMed] [Google Scholar]

[CR31] Chow ML, Pramparo T, Winn ME, Barnes CC, Li H-R, Weiss L, Fan J-B, Murray S, April C, Belinson H, Fu X-D, Wynshaw-Boris A, Schork NJ, Courchesne E. Age-dependent brain gene expression and copy number anomalies in autism suggest distinct pathological processes at young versus mature ages. PLoS Genet. 2012;8:e1002592. doi: 10.1371/journal.pgen.1002592. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR32] Colombo J, Brez CC, Curtindale LM (2012) Infant perception and cognition. In: Lerner RM, Easterbrooks MA, Mistry J (eds) Handbook of psychology, vol 6: developmental psychology. Wiley, Hoboken, NJ, pp 61–89

[CR33] Corbetta M, Shulman GL. Control of goal-directed and stimulus-driven attention in the brain. Nat Rev Neurosci. 2002;3:201–215. doi: 10.1038/nrn755. [DOI] [PubMed] [Google Scholar]

[CR34] Courchesne E, Pierce K. Why the frontal cortex in autism might be talking only to itself: local over-connectivity but long-distance disconnection. Curr Opin Neurobiol. 2005;15:225–230. doi: 10.1016/j.conb.2005.03.001. [DOI] [PubMed] [Google Scholar]

[CR35] Courchesne E, Redcay E, Morgan JT, Kennedy DP. Autism at the beginning: microstructural and growth abnormalities underlying the cognitive and behavioral phenotype of autism. Dev Psychopathol. 2005;17:577–597. doi: 10.1017/S0954579405050285. [DOI] [PubMed] [Google Scholar]

[CR36] Courchesne E, Pierce K, Schumann CM, Redcay E, Buckwalter JA, Kennedy DP, Morgan JT. Mapping early brain development in autism. Neuron. 2007;56:399–413. doi: 10.1016/j.neuron.2007.10.016. [DOI] [PubMed] [Google Scholar]

[CR37] Crespi B, Badcock C. Psychosis and autism as diametrical disorders of the social brain. Behav Brain Sci. 2008;31:241–320. doi: 10.1017/S0140525X08004214. [DOI] [PubMed] [Google Scholar]

[CR38] Currenti SA. Understanding and determining the etiology of autism. Cell Mol Neurobiol. 2010;30:161–171. doi: 10.1007/s10571-009-9453-8. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR39] Damoiseaux JS, Rombouts SARB, Barkhof F, Scheltens P, Stam CJ, Smith SM, Beckmann CF. Consistent resting-state networks across healthy subjects. Proc Natl Acad Sci USA. 2006;103:13848–13853. doi: 10.1073/pnas.0601417103. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR40] Dehaene S, Changeux JP. Neural mechanisms for access to consciousness. In: Gazzaniga MS, editor. The cognitive neurosciences. 3. Cambridge, MA: MIT Press; 2004. pp. 1145–1157. [Google Scholar]

[CR41] Dehaene S, Changeux JP. Experimental and theoretical approaches to conscious processing. Neuron. 2011;70:200–227. doi: 10.1016/j.neuron.2011.03.018. [DOI] [PubMed] [Google Scholar]

[CR42] Dehaene S, Naccache L. Towards a cognitive neuroscience of consciousness: basic evidence and a workspace framework. Cognition. 2001;79:1–37. doi: 10.1016/S0010-0277(00)00123-2. [DOI] [PubMed] [Google Scholar]

[CR43] Dehaene S, Charles L, King J-R, Marti S. Toward a computational theory of conscious processing. Curr Opin Neurobiol. 2014;25:76–84. doi: 10.1016/j.conb.2013.12.005. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR44] Dehaene-Lambertz G, Spelke ES. The infancy of the human brain. Neuron. 2015;88:93–109. doi: 10.1016/j.neuron.2015.09.026. [DOI] [PubMed] [Google Scholar]

[CR45] DeRamus TP, Black BS, Pennick MR, Kana RK. Enhanced parietal cortex activation during location detection in children with autism. Neurodev Disord. 2014;6:37. doi: 10.1186/1866-1955-6-37. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR46] Di Martino A, Fair DA, Kelly C, Satterthwaite TD, Castellanos FX, Thomason ME, Craddock RC, Luna B, Leventhal BL, Zhuo X-N, Milham MP. Unraveling the miswired connectome: a developmental perspective. Neuron. 2014;83:1335–1353. doi: 10.1016/j.neuron.2014.08.050. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR47] Dinstein I, Pierce K, Eyler L, Solso S, Malach R, Behrmann M, Courchesne E. Disrupted neural synchronization in toddlers with autism. Neuron. 2011;70:1218–1225. doi: 10.1016/j.neuron.2011.04.018. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR48] Doria V, Beckmann CF, Arichi T, Merchant N, Groppo M, Turkheimer FE, Counsell SJ, Murgasova M, Aljabar P, Nunes RG, Larkman DJ, Rees G, Edwards AD. Emergence of resting state networks in the preterm human brain. Proc Natl Acad Sci USA. 2010;107:20015–20020. doi: 10.1073/pnas.1007921107. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR49] Elsabbagh M, Fernandes J, Webb SJ, Dawson G, Charman T, Johnson MH, The British Autism Study of Infant Siblings Team Disengagement of visual attention in infancy is associated with emerging autism in toddlerhood. Biol Psychol. 2013;74:189–194. doi: 10.1016/j.biopsych.2012.11.030. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR50] Elton A, Di Martino A, Hazlett HC, Gao W. Neural connectivity evidence for a categorical-dimensional hybrid model of autism spectrum disorder. Biol Psychiatry. 2016;80:120–128. doi: 10.1016/j.biopsych.2015.10.020. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR51] Fair DA, Cohen AL, Dosenbach NUF, Church JA, Miezin FM, Barch DM, Raichle MA, Petersen SE, Schlaggar BL. The maturing architecture of the brain’s default network. Proc Natl Acad Sci USA. 2008;105:4028–4032. doi: 10.1073/pnas.0800376105. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR52] Fernandino L, Binder JR, Desai RH, Pendl SL, Humphries CJ, Gross WL, Conant LL, Seidenberg MS. Concept representation reflects multimodal abstraction: a framework for embodied semantics. Cortex. 2015 doi: 10.1093/cercor/bhv020. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR53] Fields C. Trajectory recognition as the basis for object individuation: a functional model of object file instantiation and object-token encoding. Front Psychol. 2011;2:49. doi: 10.3389/fpsyg.2011.00049. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR54] Fields C. Do autism spectrum disorders involve a generalized object categorization and identification dysfunction? Med Hypotheses. 2012;79:344–351. doi: 10.1016/j.mehy.2012.05.032. [DOI] [PubMed] [Google Scholar]

[CR55] Fields C. The very same thing: extending the object token concept to incorporate causal constraints on individual identity. Adv Cogn Psychol. 2012;8:234–247. doi: 10.5709/acp-0119-8. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR56] Fields C. The principle of persistence, Leibniz’s law, and the computational task of object re-identification. Hum Dev. 2013;56:147–166. doi: 10.1159/000350206. [DOI] [Google Scholar]

[CR57] Fields C. Motion, identity and the bias toward agency. Front Hum Neurosci. 2014;8:597. doi: 10.3389/fnhum.2014.00597. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR58] Fornito A, Zalesky A, Bassett DS, Meunier D, Ellison-Wright I, Yücel M, Wood SJ, Shaw K, O’Connor J, Nertney D, Mowry BJ, Pantelis C, Bullmore ET. Genetic influences on cost-efficient organization of human cortical functional networks. J Neurosci. 2011;31:3261–3270. doi: 10.1523/JNEUROSCI.4858-10.2011. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR59] Franklin S, Ramamurthy U, D’Mello SK, McCauley L, Negatu A, Silva LR, Datla V (2007) LIDA: a computational model of global workspace theory and developmental learning. In: Proceedings of the AAAI fall symposium on artificial intelligence and consciousness. AAAI, Menlo Park, CA, pp 61–66

[CR60] Friston KJ. The free-energy principle: a unified brain theory? Nat Rev Neurosci. 2010;11:127–138. doi: 10.1038/nrn2787. [DOI] [PubMed] [Google Scholar]

[CR61] Friston KJ. Functional and effective connectivity: a review. Brain Connect. 2011;1:13–36. doi: 10.1089/brain.2011.0008. [DOI] [PubMed] [Google Scholar]

[CR62] Friston K, Lawson R, Frith C. On hyperpriors and hypopriors: comment on Pellicano and Burr. Trends Cogn Sci. 2012;17:1. doi: 10.1016/j.tics.2012.11.003. [DOI] [PubMed] [Google Scholar]

[CR63] Gallivan JP, Culham JC. Neural coding within human brain areas involved in actions. Curr Opin Neurobiol. 2015;33:141–149. doi: 10.1016/j.conb.2015.03.012. [DOI] [PubMed] [Google Scholar]

[CR64] Gao W, Gilmore JH, Shen D, Smith JK, Zhu H, Lin W. The synchronization within and interaction between the default and dorsal attention networks in early infancy. Cereb Cortex. 2013;23:594–603. doi: 10.1093/cercor/bhs043. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR65] Gao W, Alcauter S, Smith JK, Gilmore JH, Lin W. Development of human brain cortical network architecture during infancy. Brain Struct Funct. 2015;220:1173–1186. doi: 10.1007/s00429-014-0710-3. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR66] Gardener H, Spiegelman D, Buka HL. Perinatal and neonatal risk factors for autism: a comprehensive meta-analysis. Pediatrics. 2011;128:344–355. doi: 10.1542/peds.2010-1036. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR67] Geschwind DH, Levitt S. Autism spectrum disorders: developmental disconnection syndromes. Curr Opin Neurobiol. 2007;17:103–111. doi: 10.1016/j.conb.2007.01.009. [DOI] [PubMed] [Google Scholar]

[CR68] Geschwind DH. Autism: many genes, common pathways? Cell. 2008;135:391–395. doi: 10.1016/j.cell.2008.10.016. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR69] Geschwind DH. Advances in autism. Annu Rev Med. 2009;60:367–380. doi: 10.1146/annurev.med.60.053107.121225. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR70] Geschwind DH, Flint J. Genetics and genomics of psychiatric disease. Science. 2015;349:1489–1494. doi: 10.1126/science.aaa8954. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR71] Gibson JJ. The ecological approach to visual perception. Boston: Houghton Mifflin; 1979. [Google Scholar]

[CR72] Glazebrook JF, Wallace R. Small worlds and red queens in the global workspace: an information-theoretic approach. Cogn Syst Res. 2009;10:333–365. doi: 10.1016/j.cogsys.2009.01.002. [DOI] [Google Scholar]

[CR73] Glazebrook JF, Wallace R. Pathologies in functional connectivity, feedback control and robustness: a global workspace perspective on autism spectrum disorders. Cogn Proc. 2015;16:1–16. doi: 10.1007/s10339-014-0636-y. [DOI] [PubMed] [Google Scholar]

[CR74] Gómez CM, Flores A. A neurophysiological evaluation of a cognitive cycle in humans. Neurosci Biobehav Rev. 2011;35:452–461. doi: 10.1016/j.neubiorev.2010.05.005. [DOI] [PubMed] [Google Scholar]

[CR75] Gomot M, Wicker B. A challenging, unpredictable world for people with autism spectrum disorder. Int J Psychophysiol. 2012;83:240–247. doi: 10.1016/j.ijpsycho.2011.09.017. [DOI] [PubMed] [Google Scholar]

[CR76] Gopnik A. Scientific thinking in young children: theoretical advances, empirical research, and policy implications. Science. 2012;337:1623–1627. doi: 10.1126/science.1223416. [DOI] [PubMed] [Google Scholar]

[CR77] Gopnik A, Wellman HM. Reconstructing constructivism: causal models, Bayesian learning mechanisms and the theory theory. Psychol Bull. 2012;138:1085–1108. doi: 10.1037/a0028044. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR78] Gottlieb J, Oudeyer P-Y, Lopes M, Baranes A. Information-seeking, curiosity, and attention: computational and neural mechanisms. Trends Cogn Sci. 2013;17:585–593. doi: 10.1016/j.tics.2013.09.001. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR79] Graziano MSA. Speculations of the evolution of awareness. J Cogn Neurosci. 2014;26:1300–1304. doi: 10.1162/jocn_a_00623. [DOI] [PubMed] [Google Scholar]

[CR80] Grossberg S, Seidman D. Neural dynamics of autistic behaviors: cognitive, emotional, and timing substrates. Psychol Rev. 2006;113:483–525. doi: 10.1037/0033-295X.113.3.483. [DOI] [PubMed] [Google Scholar]

[CR81] Hahamy A, Behrmann M, Malach R. The idiosyncratic brain: distortion of spontaneous connectivity patterns in autism spectrum disorder. Nat Neurosci. 2015;18:302–309. doi: 10.1038/nn.3919. [DOI] [PubMed] [Google Scholar]

[CR82] Happé F, Frith U. The weak coherence account: detail-focused cognitive style in autism spectrum disorders. J Autism Dev Disord. 2006;36:5–25. doi: 10.1007/s10803-005-0039-0. [DOI] [PubMed] [Google Scholar]

[CR83] Harms MB, Martin A, Wallace GL. Facial emotion recognition in autism spectrum disorders: a review of behavioral and neuroimaging studies. Neuropsychol Rev. 2010;20:290–322. doi: 10.1007/s11065-010-9138-6. [DOI] [PubMed] [Google Scholar]

[CR84] Hellendoorn A, Wijnroks L, Leseman PPM. Unraveling the nature of autism: finding order amid change. Front Psychol. 2015;6:359. doi: 10.3389/fpsyg.2015.00359. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR85] Hoehl S. The development of category specificity in infancy—what can we learn from electrophysiology? Neuropsychologia. 2016;83:114–122. doi: 10.1016/j.neuropsychologia.2015.08.021. [DOI] [PubMed] [Google Scholar]

[CR86] Hoelh S, Peykarjou S. The early development of face processing—what makes faces special? Neurosci Bull. 2012;28:765–788. doi: 10.1007/s12264-012-1280-0. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR87] Hoffman D, Prakash C. Objects of consciousness. Front Psychol. 2014;5:577. doi: 10.3389/fpsyg.2014.00577. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR88] Huang H, Shu N, Mishra V, Jeon T, Chalak L, Wang ZJ, Rollins N, Gong G, Cheng H, Peng Y, Dong Q, He Y. Development of human brain structural networks through infancy and childhood. Cereb Cortex. 2015;25:1389–1404. doi: 10.1093/cercor/bht335. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR89] Iacoboni M, Dapretto M. The mirror neuron system and the consequences of its dysfunction. Nat Rev Neurosci. 2006;7:942–951. doi: 10.1038/nrn2024. [DOI] [PubMed] [Google Scholar]

[CR90] Ishibashi R, Pobric G, Saito S, Lambon Ralph MA. The neural network for tool-related cognition: an activation likelihood estimation meta-analysis of 70 neuroimaging contrasts. Cogn Neuropsychol. 2016;33:241–256. doi: 10.1080/02643294.2016.1188798. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR91] Jarman N, Trengove C, Steur E, Tyukin I, van Leeuwen C. Spatially constrained adaptive rewiring in cortical networks creates spatially modular small world architectures. Cogn Neurodyn. 2014;8:479–497. doi: 10.1007/s11571-014-9288-y. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR92] Jeste SS. The neurology of autism spectrum disorders. Curr Opin Neurol. 2011;24:132–139. doi: 10.1097/WCO.0b013e3283446450. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR93] Johnson MH. Interactive specialization: a domain-general framework for human functional brain development? Dev Cogn Neurosci. 2011;1:7–21. doi: 10.1016/j.dcn.2010.07.003. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR94] Just M, Keller T, Malave V, Kana R, Varma S. Autism as a neural disorder: a theory of frontal-posterior underconnectivity. Neurosci Biobehav Rev. 2012;36:1292–1313. doi: 10.1016/j.neubiorev.2012.02.007. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR95] Kaiser MD, Pelphrey KA. Disrupted action perception in autism: behavioral evidence, neuroendophenotypes, and diagnostic utility. Dev Cogn Neurosci. 2012;2:25–35. doi: 10.1016/j.dcn.2011.05.005. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR96] Keehn B, Wagner JB, Tager-Flusberg H, Nelson CA. Functional connectivity in the first year of life in infants at-risk for autism: a preliminary near-infrared spectroscopy study. Front Human Neurosci. 2013;7:444. doi: 10.3389/fnhum.2013.00444. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR97] Kennedy DP, Semendeferi K, Courchesne E. No reduction of spindle neuron number in frontoinsular cortex in autism. Brain Cogn. 2007;64:124–129. doi: 10.1016/j.bandc.2007.01.007. [DOI] [PubMed] [Google Scholar]

[CR98] Kidd C, Piantadosi ST, Aslin RN. The Goldilocks effect: human infants allocate attention to visual sequences that are neither too simple nor too complex. PLoS ONE. 2012;7:e36399. doi: 10.1371/journal.pone.0036399. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR99] Kiefer M, Pulvermüller F. Conceptual representations in mind and brain: theoretical developments, current evidence and future directions. Cortex. 2012;48:805–825. doi: 10.1016/j.cortex.2011.04.006. [DOI] [PubMed] [Google Scholar]

[CR100] Kita EM, Scott EK, Goodhill GJ. The influence of activity on axon pathfinding in the optic tectum. Dev Neurobiol. 2015;75:608–620. doi: 10.1002/dneu.22262. [DOI] [PubMed] [Google Scholar]

[CR101] Knill DC, Pouget A. The Bayesian brain: the role of uncertainty in neural coding and computation. Trends Neurosci. 2004;27:712–719. doi: 10.1016/j.tins.2004.10.007. [DOI] [PubMed] [Google Scholar]

[CR102] Latora V, Marchiori M. Efficient behavior of small-world networks. Phys Rev Lett. 2001;87:198701. doi: 10.1103/PhysRevLett.87.198701. [DOI] [PubMed] [Google Scholar]

[CR103] Lawson RP, Rees G, Friston KJ. An aberrant precision account of autism. Front Hum Neurosci. 2014;8:302. doi: 10.3389/fnhum.2014.00302. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR104] Leonard H, Dixon G, Whitehouse AJO, et al. Unpacking the complex nature of the autism epidemic. Res Autism Spectrum Disord. 2010;4:548–554. doi: 10.1016/j.rasd.2010.01.003. [DOI] [Google Scholar]

[CR105] Lewis JW. Cortical networks related to human use of tools. Neuroscientist. 2006;12:211–231. doi: 10.1177/1073858406288327. [DOI] [PubMed] [Google Scholar]

[CR106] Lewis JD, Evans AC, Pruett JR, Botterton K, Zwaigenbaum L, Estes A, Gerig G, Collins L, Kostopoulos P, McKinstry R, Dages S, Paterson S, Schultz RT, Styner M, Hazlett H, Piven J, (IBIS Network) Network inefficiencies in autism spectrum disorder at 24 months. Transl Psychiatry. 2014;4:e388. doi: 10.1038/tp.2014.24. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR107] Lombardo MV, Pierce K, Eyler LT, Barnes CC, Ahrens-Barbeau C, Solso S, Campbell K, Courchesne E. Different functional neural substrates for good and poor language outcome in autism. Neuron. 2015;86:567–577. doi: 10.1016/j.neuron.2015.03.023. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR108] Luo Y. Three-month-old infants attribute goals to a non-human agent. Dev Sci. 2011;14:453–460. doi: 10.1111/j.1467-7687.2010.00995.x. [DOI] [PubMed] [Google Scholar]

[CR109] Luo Y, Baillargeon R. Toward a mentalistic account of early psychological reasoning. Curr Dir Psychol Sci. 2010;19:301–307. doi: 10.1177/0963721410386679. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR110] Markram H, Rinaldi T, Markram K. The intense world syndrome—an alternative hypothesis for autism. Front Neurosci. 2007;1:77–96. doi: 10.3389/neuro.01.1.1.006.2007. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR111] Matson JL, Kozlowski AM. The increasing prevalence of autism spectrum disorders. Res Autism Spectrum Disord. 2011;5:418–425. doi: 10.1016/j.rasd.2010.06.004. [DOI] [Google Scholar]

[CR112] Maximo JO, Cadena EJ, Kana RK. The implications of brain connectivity in the neuropsychology of autism. Neuropsychol Rev. 2014;24:16–31. doi: 10.1007/s11065-014-9250-0. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR113] McCall R, Franklin S (2013) Cortical learning algorithms with predictive coding for a systems-level cognitive architecture. In: Proceedings of the second annual conference on advances in cognitive systems. Cognitive Systems Foundation, pp 149–166

[CR114] McFadden K, Minshew NJ. Evidence for dysregulation of axonal growth and guidance in the etiology of ASD. Front Hum Neurosci. 2013;7:671. doi: 10.3389/fnhum.2013.00671. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR115] Miles JH. Autism subgroups from a medical genetics perspective. In: Amaral DG, Dawson G, Geschwind DH, editors. Autism spectrum disorders. New York: Oxford University Press; 2011. pp. 705–721. [Google Scholar]

[CR116] Minshew NJ, Williams DL. The new neurobiology of autism. Arch Neurol. 2007;64:945–950. doi: 10.1001/archneur.64.7.945. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR117] Müller R-A, Shih P, Keehn B, Deyoe JR, Leyden KN, Shukla DK. Underconnected, but how? A survey of functional connectivity MRI studies in autism spectrum disorders. Cereb Cortex. 2011;21:2233–2243. doi: 10.1093/cercor/bhq296. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR118] Munz M, Gobert D, Schohl A, Poquérusse J, Podgorski K, Spratt P, Ruthazer ES. Rapid Hebbian axonal remodeling mediated by visual stimulation. Science. 2014;344:904–909. doi: 10.1126/science.1251593. [DOI] [PubMed] [Google Scholar]

[CR119] Needham A, Dueker G, Lockhead G. Infants’ formation and use of categories to segregate objects. Cognition. 2005;94:215–240. doi: 10.1016/j.cognition.2004.02.002. [DOI] [PubMed] [Google Scholar]

[CR120] Newman MEJ. The structure and function of complex networks. SIAM Rev. 2003;45:167–256. doi: 10.1137/S003614450342480. [DOI] [Google Scholar]

[CR121] Oberman LM, Ramachandran VS. The simulating social mind: the role of the mirror neuron system and simulation in the social and communicative deficits of autism spectrum disorders. Psychol Bull. 2007;133:310–327. doi: 10.1037/0033-2909.133.2.310. [DOI] [PubMed] [Google Scholar]

[CR122] Osiurak F, Jarry C, Lesourd M, Baumard J, Le Gall D. Mechanical problem-solving strategies in left-brain damaged patients and apraxia of tool use. Neuropsychology. 2013;51:1964–1972. doi: 10.1016/j.neuropsychologia.2013.06.017. [DOI] [PubMed] [Google Scholar]

[CR123] Oudeyer P-Y, Baranes A, Kaplan F. Intrinsically motivated learning of real world sensorimotor skills with developmental constraints. In: Baldassarre G, Mirolli M, editors. Intrinsically motivated learning in natural and artificial systems. Berlin: Springer; 2013. pp. 303–365. [Google Scholar]

[CR124] Parikshak NN, Luo R, Zhang A, Won H, Lowe JK, Chandran V, Horvath S, Geschwind DH. Integrative functional genomic analyses implicate specific molecular pathways and circuits in autism. Cell. 2013;155:1008–1021. doi: 10.1016/j.cell.2013.10.031. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR125] Park H-J, Friston K. Structural and functional brain networks: from connections to cognition. Science. 2013;342:1238411-1–1238411-8. doi: 10.1126/science.1238411. [DOI] [PubMed] [Google Scholar]

[CR126] Peca J, Feng G. Cellular and synaptic network defects in autism. Curr Opin Neurol. 2012;22:866–872. doi: 10.1016/j.conb.2012.02.015. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR127] Pellicano E, Burr D. When the world becomes `too real’: a Bayesian explanation of autistic perception. Trends Cogn Sci. 2012;16:504–510. doi: 10.1016/j.tics.2012.08.009. [DOI] [PubMed] [Google Scholar]

[CR128] Pelphrey KA, Shultz S, Hudac CM, Vander Wyk BC. Constraining heterogeneity: the social brain and its development in autism spectrum disorder. J Child Psychol Psychiatry. 2011;52:631–644. doi: 10.1111/j.1469-7610.2010.02349.x. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR129] Peters JM, Taquet M, Vega C, et al. Brain functional networks in syndromic and non-syndromic autism: a graph theoretical study of EEG connectivity. BMC Med. 2013;11:54. doi: 10.1186/1741-7015-11-54. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR130] Picci G, Gotts SJ, Scherf KS. A theoretical rut: revisiting and critically evaluating the generalized under/over-connectivity hypothesis of autism. Dev Sci. 2016;19:524–549. doi: 10.1111/desc.12467. [DOI] [PubMed] [Google Scholar]

[CR131] Pinto D, Delaby E, Merico D, et al. Convergence of genes and cellular pathways dysregulated in autism spectrum disorders. Am J Hum Genet. 2014;94:677–694. doi: 10.1016/j.ajhg.2014.03.018. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR132] Rajendran G, Mitchell P. Cognitive theories of autism. Dev Rev. 2007;27:224–260. doi: 10.1016/j.dr.2007.02.001. [DOI] [Google Scholar]

[CR133] Rakison DH, Yermoleva Y. Infant categorization. Wiley Interdiscip Rev Cogn Sci. 2010;1:894–905. doi: 10.1002/wcs.81. [DOI] [PubMed] [Google Scholar]

[CR134] Righi G, Tierney AL, Tager-Flusberg H, Nelson CA. Functional connectivity in the first year of life in infants at risk for autism spectrum disorder: an EEG study. PLoS ONE. 2014;9:e105176. doi: 10.1371/journal.pone.0105176. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR135] Rippon G, Brock J, Brown C, Boucher J. Disordered connectivity in the autistic brain: challenges for the ‘new psychophysiology’. Int J Psychophysiol. 2007;63:164–172. doi: 10.1016/j.ijpsycho.2006.03.012. [DOI] [PubMed] [Google Scholar]

[CR136] Robertson CE, Thomas C, Kravitz DJ, Wallace GL, Baron-Cohen S, Martin A, Baker CI. Global motion perception deficits in autism are reflected as early as primary visual cortex. Brain. 2014;137:2588–2599. doi: 10.1093/brain/awu189. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR137] Rochat P. Primordial sense of embodied self-unity. In: Slaughter V, Brownell CA, editors. Early development of body representations. Cambridge: Cambridge University Press; 2012. pp. 3–18. [Google Scholar]

[CR138] Rolls ET. Limbic systems for emotion and for memory, but no single limbic system. Cortex. 2015;62:119–157. doi: 10.1016/j.cortex.2013.12.005. [DOI] [PubMed] [Google Scholar]

[CR139] Rubinov M, Sporns O. Complex network measures of brain connectivity: uses and interpretations. NeuroImage. 2010;52:1059–1069. doi: 10.1016/j.neuroimage.2009.10.003. [DOI] [PubMed] [Google Scholar]

[CR140] Rugg MD, Vilberg KL. Brain networks underlying episodic memory retrieval. Curr Opin Neurobiol. 2013;23:255–260. doi: 10.1016/j.conb.2012.11.005. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR141] Santos M, Uppal N, Butti C, Wicinski B, Schmeidler J, Giannakopoulos P, Heinsen H, Schmitz C, Hof PR. von Economo neurons in autism: a stereologic study of the frontoinsular cortex in children. Brain Res. 2011;1380:206–217. doi: 10.1016/j.brainres.2010.08.067. [DOI] [PubMed] [Google Scholar]

[CR142] Schlesinger M, Amso D, Johnson SP, Hantehzadeh N, Gupta L (2012) Using the iCub simulator to study perceptual development: a case study. In: Proceedings of the 2012 IEEE international conference on development and learning and epigenetic robotics (ICDL). IEEE, pp 1–6. doi:10.1109/DevLrn.2012.6400866

[CR143] Shanahan M. The brain’s connective core and its role in animal cognition. Philos Trans R Soc B. 2012;367:2704–2714. doi: 10.1098/rstb.2012.0128. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR144] Shih P, Shen M, Öttl B, Keehn B, Gaffrey MS, Müller R-A. Atypical network connectivity for imitation in autism spectrum disorder. Neuropsychologia. 2010;48:2931–2939. doi: 10.1016/j.neuropsychologia.2010.05.035. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR145] Shipp S, Adams RA, Friston KJ. Reflections on agranular architecture: predictive coding in the motor cortex. Trends Neurosci. 2013;36:706–716. doi: 10.1016/j.tins.2013.09.004. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR146] Simion F, Di Giorgio E, Leo I, Bardi L. The processing of social stimuli in early infancy: from faces to biological motion. Prog Brain Res. 2011;189:173–193. doi: 10.1016/B978-0-444-53884-0.00024-5. [DOI] [PubMed] [Google Scholar]

[CR147] Simms ML, Kemper TL, Timbie CM, Bauman ML, Blatt GJ. The anterior cingulate cortex in autism: heterogeneity of qualitative and quantitative cytoarchitectonic features suggests possible subgroups. Acta Neuropathol. 2009;118:673–684. doi: 10.1007/s00401-009-0568-2. [DOI] [PubMed] [Google Scholar]

[CR148] Sporns O. Network attributes for segregation and integration in the human brain. Curr Opin Neurobiol. 2013;23:162–171. doi: 10.1016/j.conb.2012.11.015. [DOI] [PubMed] [Google Scholar]

[CR149] Sporns O, Honey CJ. Small worlds inside big brains. Proc Natl Acad Sci USA. 2006;111:15220–15225. doi: 10.1073/pnas.0609523103. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR150] Sporns O, Zwi JD. The small world of the cerebral cortex. Neuroinformatics. 2004;2:145–162. doi: 10.1385/NI:2:2:145. [DOI] [PubMed] [Google Scholar]

[CR151] Sporns O, Tononi G, Edelman GM. Theoretical neuroanatomy and the connectivity of the cerebral cortex. Behav Brain Res. 2002;135:69–74. doi: 10.1016/S0166-4328(02)00157-2. [DOI] [PubMed] [Google Scholar]

[CR152] Stahl AE, Feigenson L. Observing the unexpected enhances infants’ learning and exploration. Science. 2015;348:91–94. doi: 10.1126/science.aaa3799. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR153] Sunderland A, Wilkins L, Dineen R, Dawson SE. Tool-use and the left hemisphere: what is lost in ideomotor apraxia? Brain Cogn. 2013;81:183–192. doi: 10.1016/j.bandc.2012.10.008. [DOI] [PubMed] [Google Scholar]

[CR154] Tanaka JW, Sung A. The “eye avoidance” hypothesis of autism face processing. J Autism Dev Disord. 2016;46:1538–1552. doi: 10.1007/s10803-013-1976-7. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR155] Tononi G, Koch C. Consciousness: here, there and everywhere? Philos Trans R Soc B. 2015;370:20140167. doi: 10.1098/rstb.2014.0167. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR156] Toulmin H, Beckmann CF, O’Muircheartaigh J, Ball G, Nongena P, Makropoulos A, Ederies A, Counsell SJ, Kennea N, Arichi T, Tusor N, Rutherford MA, Azzopardi D, Gonzalez-Cinca N, Hajnal JV, Edwards AD. Specialization and integration of functional thalamocortical connectivity in the human infant. Proc Natl Acad Sci USA. 2015;112:6485–6490. doi: 10.1073/pnas.1422638112. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR157] Tyszka JM, Kennedy DP, Paul LK, Adolphs R. Largely typical patterns of resting-state functional connectivity in high-functioning adults with autism. Cereb Cortex. 2014;24:1894–1905. doi: 10.1093/cercor/bht040. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR158] Uddin LQ. Salience processing and insular cortical function and dysfunction. Nat Rev Neurosci. 2015;16:55–61. doi: 10.1038/nrn3857. [DOI] [PubMed] [Google Scholar]

[CR159] Uddin LQ, Supekar K, Lynch CJ, Khouzam A, Phillips J, Feinsten C, Ryall S, Menon V. Salience network-based classification and prediction of symptom severity in children with autism. JAMA Psychiatry. 2013;70:869–879. doi: 10.1001/jamapsychiatry.2013.104. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR160] Uddin LQ, Supekar K, Menon V. Reconceptualizing functional brain connectivity in autism from a developmental perspective. Front Hum Neurosci. 2013;7:458. doi: 10.3389/fnhum.2013.00458. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR161] Vaccarino FM, Grigorenko EL, Smith KM, Stevens H. Regulation of cerebral cortical size and neuron number by fibroblast growth factors: implications for autism. J. Autism Dev Disord. 2009;39:511–520. doi: 10.1007/s10803-008-0653-8. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR162] van Boxtel JJA, Lu H. A predictive coding perspective on autism spectrum disorders. Front Psychol. 2013;4:19. doi: 10.3389/fpsyg.2013.00019. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR163] Van de Cruys S, Evers K, Van der Hallen R, Van Eylen L, Boets B, de-Wit L, Wagemans J. Precise minds in uncertain worlds: predictive coding in autism. Psychol Rev. 2014;121:649–675. doi: 10.1037/a0037665. [DOI] [PubMed] [Google Scholar]

[CR164] van den Heuvel MP, Sporns O. Rich-club organization of the human connectome. J Neurosci. 2011;31:15775–15786. doi: 10.1523/JNEUROSCI.3539-11.2011. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR165] van Leeuwen C. What makes you think you are conscious? An agnosticist manifesto. Front Hum Neurosci. 2015;9:170. doi: 10.3389/fnhum.2015.00170. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR166] Vértes PE, Bullmore ET. Annual research review: growth connectomics—the organization and reorganization of brain networks during normal and abnormal development. J Child Psychol Psychiatry. 2015;56:299–320. doi: 10.1111/jcpp.12365. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR167] Vingerhoets G. Contribution of the posterior parietal cortex in reaching, grasping, and using objects and tools. Front Psychol. 2014;5:151. doi: 10.3389/fpsyg.2014.00151. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR168] Vissers ME, Cohen MX, Geurts HM. Brain connectivity and high functioning autism: a promising path of research that needs refined models, methodological convergence, and strong behavioral links. Neurosci Biobehav Rev. 2012;36:604–625. doi: 10.1016/j.neubiorev.2011.09.003. [DOI] [PubMed] [Google Scholar]

[CR169] Volkmar F, Chawarska K, Klin A. Autism in infancy and early childhood. Annu Rev Psychol. 2005;56:315–336. doi: 10.1146/annurev.psych.56.091103.070159. [DOI] [PubMed] [Google Scholar]

[CR170] von Hofsten C. Action in development. Dev Sci. 2007;10:54–60. doi: 10.1111/j.1467-7687.2007.00564.x. [DOI] [PubMed] [Google Scholar]

[CR171] Vossel S, Geng JJ, Fink GR. Dorsal and ventral attention systems: distinct neural circuits but collaborative roles. Neurosci. 2014;20:150–159. doi: 10.1177/1073858413494269. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR172] Wallace R. Consciousness: a mathematical treatment of the global neuronal workspace model. New York: Springer; 2005. [Google Scholar]

[CR173] Wass S. Distortions and disconnections: disrupted brain connectivity in autism. Brain Cogn. 2011;75:18–28. doi: 10.1016/j.bandc.2010.10.005. [DOI] [PubMed] [Google Scholar]

[CR174] Watts DJ, Strogatz SH. Collective dynamics of ‘small-world’ networks. Nature. 1998;393:440–442. doi: 10.1038/30918. [DOI] [PubMed] [Google Scholar]

[CR175] Weigelt S, Koldewyn K, Kanwisher N. Face identity recognition in autism spectrum disorders: a review of behavioral studies. Neurosci Biobehav Rev. 2012;36:1060–1084. doi: 10.1016/j.neubiorev.2011.12.008. [DOI] [PubMed] [Google Scholar]

[CR176] Westermann G, Mareschal D. From perceptual to language-mediated categorization. Philos Trans R Soc B. 2012;369:20120391. doi: 10.1098/rstb.2012.0391. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR177] Williams JH, Whiten A, Suddendorf T, Perrett DI. Imitation, mirror neurons and autism. Neurosci Biobehav Rev. 2001;25:287–295. doi: 10.1016/S0149-7634(01)00014-8. [DOI] [PubMed] [Google Scholar]

[CR178] Willsey AJ, Sanders SJ, Li M, et al. Coexpression networks implicate human midfetal deep cortical projection neurons in the pathogenesis of autism. Cell. 2013;155:997–1007. doi: 10.1016/j.cell.2013.10.020. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR179] Yap P-T, Fan Y, Chen Y, Gilmore JH, Lin W, Shen D. Development trends of white matter connectivity in the first years of life. PLoS ONE. 2011;6:e24678. doi: 10.1371/journal.pone.0024678. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR180] Yeo BTT, Krienen FM, Sepulcre J, Sabuncu MR, Lashkari D, Hollinshead M, Roffman JL, Smoller JW, Zöllei L, Polimeni JR, Fischl B, Liu H, Buckner RL. The organization of the human cerebral cortex estimated by intrinsic functional connectivity. J Neurophysiol. 2011;106:1125–1165. doi: 10.1152/jn.00338.2011. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR181] Zimmer HD, Ecker UKD. Remembering perceptual features unequally bound in object and episodic tokens: neural mechanisms and their electrophysiological correlates. Neurosci Biobehav Rev. 2010;34:1066–1079. doi: 10.1016/j.neubiorev.2010.01.014. [DOI] [PubMed] [Google Scholar]

[CR182] Zmigrod S, de Sonneville LMJ, Colzato LS, Swaab H, Hommel B. Cognitive control of feature bindings: evidence from children with autistic spectrum disorder. Psychol Res. 2013;77:147–154. doi: 10.1007/s00426-011-0399-3. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR183] Zwaigenbaum L, Young GS, Stone WL, Dobkins K, Ozonoff S, Brian J, Bryson SE, Carver LJ, Hutman T, Iverson JM, Landa RJ, Messinger D. Early head growth in infants at risk of autism: a baby siblings research consortium study. J Am Acad Child Adolesc Psychiatry. 2014;53:1053–1062. doi: 10.1016/j.jaac.2014.07.007. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Disrupted development and imbalanced function in the global neuronal workspace: a positive-feedback mechanism for the emergence of ASD in early infancy

Chris Fields

James F Glazebrook

Abstract

Introduction

Fig. 1.

Background: the GNW as an SWN of SWNs

Fig. 2.

Fig. 3.

Development of the GNW

Phenomenology of the infant world

Functional connectivity in the infant GNW

A graphic notation for visualizing typical and variant GNW development

Fig. 4.

Perturbations of GNW development

General versus localized developmental GNW disruptions

Perturbations of salience

Perturbations of categorization

Consequences of salience and categorization perturbations for the default and executive systems

Five recent models of ASD from a GNW perspective

Motion-based categories (Fields 2012a)

Fig. 5.

Attenuated expectations (Pellicano and Burr 2012)

Fig. 6.

Overly-precise perceptions (Lawson et al. 2014; Van de Cruys et al. 2014)

Fig. 7.

Disrupted invariance detection (Hellendoorn et al. 2015)

Conclusion

Acknowledgements

Compliance with ethical standards

Conflict of interest

Contributor Information

References

ACTIONS

PERMALINK

RESOURCES

Cite

Add to Collections

PERMALINK

Disrupted development and imbalanced function in the global neuronal workspace: a positive-feedback mechanism for the emergence of ASD in early infancy

Chris Fields

James F Glazebrook

Abstract

Introduction

Fig. 1.

Background: the GNW as an SWN of SWNs

Fig. 2.

Fig. 3.

Development of the GNW

Phenomenology of the infant world

Functional connectivity in the infant GNW

A graphic notation for visualizing typical and variant GNW development

Fig. 4.

Perturbations of GNW development

General versus localized developmental GNW disruptions

Perturbations of salience

Perturbations of categorization

Consequences of salience and categorization perturbations for the default and executive systems

Five recent models of ASD from a GNW perspective

Motion-based categories (Fields 2012a)

Fig. 5.

Attenuated expectations (Pellicano and Burr 2012)

Fig. 6.

Overly-precise perceptions (Lawson et al. 2014; Van de Cruys et al. 2014)

Fig. 7.

Disrupted invariance detection (Hellendoorn et al. 2015)

Conclusion

Acknowledgements

Compliance with ethical standards

Conflict of interest

Contributor Information

References

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases