Situated minds: conceptual and emotional blending in neurodegeneration and beyond

Abstract

This scientific commentary refers to ‘When affect overlaps with concept: emotion recognition in semantic variant of primary progressive aphasia’, by Bertoux et al. (doi:10.1093/brain/awaa313).

This scientific commentary refers to ‘When affect overlaps with concept: emotion recognition in semantic variant of primary progressive aphasia’, by Bertoux et al. (doi:10.1093/brain/awaa313).

In each passing moment, our brains engage with multimodal stimuli that are both internal and external in nature. Imagine you are having an emotionally involved discussion with your partner. Focus on your own internal process while you interact with them. You are recruiting—predominantly on a subconscious level—basic cognitive skills (e.g. visuoauditory attention, sensorimotor processing, emotion recognition and interoceptive awareness) as you attempt to process the situation. Equally and concurrently, your brain is engaging multiple sources of conceptual knowledge related to past experiences. In daily interactions, affective processes are strongly intertwined, by default, with semantic (emotional and cognitive) knowledge from both experiential and declarative sources. This network of cognitive processes yields specific meaning to its beholder and allows an individual to anticipate and comprehend their socioemotional interactions in a multi-dimensional context. This view of emotion recognition ‘in the wild’ seems to be more synergistic and complex than the isolated, compartmentalized understanding of the emotion recognition process upheld by traditional neuroscientific laboratories. Even if emotion could be phenomenologically differentiated and associated with discrete brain activity, the experiential and conceptual processes dissected and isolated in neurocognitive laboratories may simply represent analytical abstractions of a holistic blending of interdependent processes (Ibanez and García, 2018). Constructionist theories are consistent with this last claim, by assuming that the emotional experience re-enacts embodied knowledge categorization (Barrett and Simmons, 2015; Adolfi et al., 2017). Prior experiences and the accumulated conceptual knowledge to which they give rise are involved in emotion recognition (Barrett and Simmons, 2015). However, these theoretical accounts seem to be in conflict with mainstream theories that assume a further compartmentalized, innate, and universal approach to affect. In this issue of Brain, Bertoux and co-workers provide evidence for the neurocognitive blending of semantic conceptual knowledge and emotion recognition in controls and patients with the semantic variant of primary progressive aphasia (svPPA) (Bertoux et al., 2020), a disease serving as a neurodegenerative lesion model (Melloni et al., 2016) of semantic knowledge impairments that has also been associated with emotional deficits (Fittipaldi et al., 2019).

This work brings novel and timely evidence of the impact of svPPA neurodegeneration on the interaction between semantic knowledge and emotion, with careful multidimensional assessments of emotions and with theoretical relevance for both svPPA and for constructionist theories of emotion. Bertoux and colleagues assessed both knowledge of concepts and recognition of basic and self-conscious facial emotional expressions presented both statically (via images) and dynamically (via videos). They assessed the relationship between valence processing and conceptual knowledge in patients and controls by combining quantitative and qualitative analysis of performance and error rates with structural (volume and connectivity) multimodal neuroimaging. Systematic deficits in conceptual knowledge and emotion recognition were revealed in patients. Importantly, these measures were significantly correlated with each other in patients and controls. Moreover, the evaluation of emotional concepts (EEC) was also impaired in patients, evidencing a deficit in taxonomic and contextual knowledge of emotions. Patients committed more valence errors, which correlated with conceptual knowledge and its anatomical structures. In addition, the authors found a reduction in the volume of frontal ventral, temporal, insular and striatal regions, together with white matter degeneration in tracts connecting frontotemporal regions, in association with both the emotional and conceptual deficits in svPPA. This fronto-insular-striato-temporal network supports the blending of valence recognition, conceptual knowledge of emotions, and emotion labelling (i.e. orbitofrontal cortex, insula, and amygdala engaged in value/affect processing, salience and interoception; and temporal regions involved in semantic processing). This consistent overlap at the behavioural and neuroimaging levels suggests an intermingling of EEC, emotion recognition, and valence. Moreover, the cognitive and anatomical overlap was not mediated by a general effect of disease severity. This cognitive and anatomical blending is theoretically consistent with cognitive neurodynamics (Park and Friston, 2013) and related principles such as downward and bottom-up emergent processing, diaschisis, and degeneracy. Additionally, this interplay remains in alignment with contextual transient, dynamic, and anticipatory stabilities—states that are difficult to quantify because of constraints associated with existing neuronal macroscales as well as environmental noise. These transient emergent states, which are not static, isolated processes, can potentially more holistically explain these interconnected cognitive processes.

In brief, the results suggest overall that conceptual knowledge and valence processing are intertwined during emotion recognition, supporting an embodied, situated, and constructionist view of emotions (Fig. 1). The results also inform recent calls for a contextual blending of different cognitive/affective processes in the clinical and neurocognitive arena. For instance, Bertoux et al. (2020) suggest that this neurocognitive blending of semantic knowledge and emotion recognition supports the social context network model (SCNM) that stresses the role of frontotemporo-insular networks involved in contextual integration and prediction of social information (Ibáñez, 2018; Ibanez and García, 2018). Interpreted within the SCNM framework (Ibáñez, 2018; Ibanez and García, 2018), the anatomical results of the present study imply that the temporal hub indexes both the experiential learning and conceptual knowledge, and the fronto-insular regions engage with the contextual update and prediction of valence and emotion recognition. Other similar interpretations of SCNM have been proposed for social cooperation (Melloni et al., 2016; Ibáñez et al., 2017). The authors also state that their results support the concept of intercognition (Ibanez and García, 2018; Ibáñez, 2019), the spontaneous blending of different cognitive and emotional processes, in the construal of emotional meanings. Their findings suggest that fragmented conceptual knowledge triggered by neurodegeneration in svPPA can have an important role in social cognition deficits. Similarly, cross-domain synergies have been shown for emotion, social cognition, and interoception (Adolfi et al., 2017). Current results from clinical neuroscience and svAPP may challenge the usual assumptions of the compartmentalized, context-free, static, and universalistic view of cognitive and affective processes (Ibanez and García, 2018).

Figure 1.

Graphic summary of conceptual and emotional blending in patients with the semantic variant of primary progressive aphasia. Left: Graphical representation of the blending of semantic knowledge (top) and facial emotion recognition (bottom), which are impaired and correlated in the semantic variant of primary progressive aphasia (svPPA, middle). Right: svPPA grey matter correlates of the evaluation of emotion concepts (EEC) and facial emotion recognition (FER) (A), and white matter correlates of FER (B). Modified from Bertoux et al. (2020). POST = posterior.

The Bertoux et al. (2020) study makes clear the need for further empirical and theoretical advances. Perspectives on the interconnected cognitive processes recruited before, during, and after the enactment of everyday behaviours present significant disparity from classical theoretical approaches. At any given instance within the temporal continuum, a single neurocognitive event (e.g. hearing a word, interacting with others) is embedded within a sequence or array of other neurocognitive processes. This perspective reveals a limitation of existing isolated, detached models; this is particularly evident when considered within the context of everyday cognition—a setting in which both external and internal variables cannot be fully controlled. A more complex and synergistic understanding of cognitive processes and their interactions and synchronicities may avoid apriorist barriers provided by putatively ‘separate’ cognitive processes. The conceptual-valence-emotion triangle highlighted by Bertoux et al. (2020) presents a case for further assessment of a more complex intercognitive perspective.

Neuroscientific research has provided correlates of compartmentalized cognitive processes such as isolated word meanings, bodiless facial recognition, context-free memory processes, passive participant responses, and intersubjective-blind interactions (Ibanez and García, 2018). Thus, the field of neuroscience has systematically produced large amounts of data related to isolated cognitive processing. These isolated analyses, however, usually cannot withstand pragmatic testing outside of the laboratory environment. An isolationist perspective of neurocognitive processing should be critically expanded to broaden the methodological approaches assessing experimental control as well as ecological validity.

Competing interests

The authors report no competing interests.

Funding

A.I. is partially supported by CONICET; FONCYT-PICT (2017-1818, 2017-1820); ANID/FONDAP (15150012); Universidad del Valle (CI 5316), Alzheimer’s Association GBHI ALZ UK-20-639295; and the Multi-Partner Consortium to Expand Dementia Research in Latin America (ReDLat), funded by the National Institutes of Aging of the National Institutes of Health (R01AG057234), the Alzheimer’s Association (SG-20-725707-ReDLat), the Rainwater Foundation, and the Global Brain Health Institute. The content is solely the responsibility of the authors and does not represent the official views of these institutions.