Disconnection syndromes of basal ganglia, thalamus, and cerebrocerebellar systems

Abstract

Disconnection syndromes were originally conceptualized as a disruption of communication between different cerebral cortical areas. Two developments mandate a re-evaluation of this notion.

First, we present a synopsis of our anatomical studies in monkey elucidating principles of organization of cerebral cortex. Efferent fibers emanate from every cortical area, and are directed with topographic precision via association fibers to ipsilateral cortical areas, commissural fibers to contralateral cerebral regions, striatal fibers to basal ganglia, and projection subcortical bundles to thalamus, brainstem and/or pontocerebellar system. We note that cortical areas can be defined by their patterns of subcortical and cortical connections. Second, we consider motor, cognitive and neuropsychiatric disorders in patients with lesions restricted to basal ganglia, thalamus, or cerebellum, and recognize that these lesions mimic deficits resulting from cortical lesions, with qualitative differences between the manifestations of lesions in functionally related areas of cortical and subcortical nodes.

We consider these findings on the basis of anatomical observations from tract tracing studies in monkey, viewing them as disconnection syndromes reflecting loss of the contribution of subcortical nodes to the distributed neural circuits. We introduce a new theoretical framework for the distributed neural circuits, based on general, and specific, principles of anatomical organization, and on the architecture of the nodes that comprise these systems. We propose that neural architecture determines function, i.e., each architectonically distinct cortical and subcortical area contributes a unique transform, or computation, to information processing; anatomically precise and segregated connections between nodes define behavior; and association fiber tracts that link cerebral cortical areas with each other enable the cross-modal integration required for evolved complex behaviors. This model enables the formulation and testing of future hypotheses in investigations using evolving magnetic resonance imaging techniques in humans, and in clinical studies in patients with cortical and subcortical lesions.

Lesion-deficit correlations in behavioral neurology have helped enhance the understanding of the human brain in connectionist and functional terms. This is exemplified by the work of Carl Wernicke (1848–1900) who sought to understand the clinical phenomenology and anatomical basis of the aphasias, and by Norman Geschwind (1926–1984) who pursued this line of reasoning in defining the clinical disconnection syndromes and their basis in neuroanatomy (see also Catani and Mesulam, 2008a, 2008b, this issue). The focus of these pioneers was on the cerebral cortex and the white matter tracts that link them, and it has become apparent that all behaviors are subserved by distributed neural systems that comprise anatomic regions, or nodes, each displaying unique architectural properties, distributed geographically throughout the nervous system, and linked anatomically and functionally in a precise and unique manner (Nauta, 1964; Luria, 1966; Geschwind, 1965a, 1965b; Pandya and Kuypers, 1969; Jones and Powell, 1970a, 1970b; Mesulam, 1981, 1990; Unger-leider and Mishkin, 1982; Goldman-Rakic, 1988).

Subcortical structures were not a consideration in the early concepts of the distributed neural circuits. These notions have required revision in the light of the description of subcortical dementia (Albert et al., 1974; Cummings and Benson, 1984), the recognition of neurobehavioral syndromes following lesions of cerebral white matter (Rao, 1996; Filley, 2001) and subcortical nuclei (Fisher, 1959; Watson and Heilman, 1979; Castaigne et al., 1981; Caplan et al., 1990), and imaging studies in psychiatric patients demonstrating abnormalities in subcortical structures (Breiter and Rauch, 1996). But the specificity of the underlying anatomical relationships and functional properties of subcortical structures have, in general, not received the same scrutiny as the cortical association areas and their association fiber pathways.

In this paper we view the clinical manifestations arising from lesions of the basal ganglia, thalamus and cerebellum as disconnection syndromes – focal disruptions of the distributed cortical and subcortical neural circuits that subserve neurologic function. We elucidate the principles of connectional neuroanatomy linking the cerebral cortex with these structures in monkey to shed light on similar hodological principles between human brains and those of non-human primates. We explore the theoretical underpinnings of the contribution of subcortical nodes to the organization of higher order behavior, and introduce a conceptual framework within which to consider the role of these subcortical nodes in the distributed circuits of the nervous system.

1. Principles of organization of fiber pathways in the cerebral cortex

Anatomical investigations of the non-human primate brain reveal that there is a consistent pattern of white matter fiber tracts that emanate from every region of the cerebral cortex (Schmahmann and Pandya, 2006). Neurons within any cortical area give rise to five distinct categories of efferent fibers. These are (1) association fibers, (2) striatal fibers, and a confluence of fibers (the “cord”) that carries (3) commissural fibers, and subcortical (projection) fibers to (4) thalamus, and (5) brainstem and/or the pontocerebellar system (Fig. 1).

Fig. 1.

Diagram (A) and schema (B) of the principles of organization of white matter fiber pathways emanating from the cerebral cortex. Long association fibers are seen end-on as the stippled area within the white matter of the gyrus. In their course, these fibers either remain confined to the white matter of the gyrus, or travel deeper in the white matter of the hemisphere. Short association fibers, or U-fibers link adjacent gyri. Neighborhood association fibers link nearby regions usually within the same lobe. Striatal fibers intermingle with the association fibers early in their course, before coursing in the subcallosal fascicle of Muratoff or in the external capsule. Cord fibers segregate into commissural fibers that arise in cortical layers II and III, and the subcortical bundle, which further divides into fibers destined for thalamus arising from cortical layer VI, and those to brainstem and spinal cord in the pontine bundle arising from cortical layer V (from Schmahmann and Pandya, 2006).

Association fibers comprise local, neighborhood and long association fibers. Local, or U-fibers, are closely apposed to the undersurface of the sixth layer, and are directed to cortical regions in the same or adjacent gyri. Neighborhood association fibers are distinct from local U-fibers and are directed to nearby cortical regions, e.g., the fibers that connect the inferior parietal lobule to the medial parietal cortex. Long association fibers are the named fiber tracts that travel within the central part of the white matter of the core of the gyrus and link distant cortical areas within the same hemisphere (Table 1).

Table 1.

Identity and location of long association fibers of the cerebral hemisphere

Long association fiber pathway	Location in cerebral white matter
Superior longitudinal fasciculus I	White matter of superior parietal lobule and superior frontal gyrus
Superior longitudinal fasciculus II	Centrum semiovale, lateral to and crossing through the corona radiata, above Sylvian fissure
Superior longitudinal fasciculus III	White matter of the parietal and frontal opercula
Arcuate fasciculus	White matter of superior temporal gyrus, and deep to upper shoulder of the Sylvian fissure, ventrally adjacent to SLF II
Middle longitudinal fasciculus	White matter of caudal inferior parietal lobule extending into white matter of the superior temporal gyrus
Extreme capsule	Between claustrum and insula caudally, and between claustrum and orbital frontal cortex rostrally
Inferior longitudinal fasciculus	Vertical limb between sagittal stratum medially and parieto-occipital and temporal cortices laterally. Horizontal component in the temporal lobe
Fronto-occipital fasciculus	Above body and head of the caudate nucleus and subcallosal fascicle of Muratoff, lateral to corpus callosum, medial to corona radiata
Uncinate fasciculus	White matter of rostral temporal lobe, limen insula, white matter of orbital and medial prefrontal cortex
Cingulum bundle	Dorsal component in white matter of the cingulate gyrus. Ventral contingent in white matter of the caudal part of parahippocampal gyrus

Corticostriatal fibers course initially with the long association fibers before separating from them, then travel within one of two major fiber bundles. Fibers in the subcallosal fasciculus of Muratoff lead mainly to the caudate nucleus and putamen. Fibers in the external capsule target the ventral part of the caudate nucleus, the putamen and claustrum (Fig. 2).

Fig. 2.

Light field photomicrograph of a coronal hemisphere of a monkey brain stained for Nissl substance, to show the locations of the Muratoff bundle (MB) coursing above the body of the caudate nucleus and below the corpus callosum, and the external capsule (EC) situated between the claustrum laterally and the putamen medially. See list of abbreviations.

The cord that contains the commissural fibers and the subcortical projection bundle emanates from each cortical region as a dense aggregation of fibers occupying the central core of the white matter of the gyrus. The fibers in the cord separate into two distinct segments. One component forms a commissural bundle that travels to the opposite hemisphere via the corpus callosum or the anterior commissure (Fig. 3). The other segment continues as a concentration of fibers, the subcortical bundle (SB), conveying the axons of infragranular neurons. Depending on its cortical origin, this subcortical bundle travels within the internal capsule (anterior or posterior limb), or the sagittal stratum. This SB segregates into a thalamic fiber bundle that travels via the thalamic peduncles to the thalamus, and a pontine fiber bundle that courses via the cerebral peduncle to the pons – the obligatory synaptic step in the feedforward limb of the cerebrocerebellar circuit. The SB also gives rise to fibers to other diencephalic and brainstem structures.

Fig. 3.

Photomicrographs of selected coronal sections of a rhesus monkey following injection of radioisotope into the frontal operculum in the precentral aspects of areas 1 and 2. A, the cord of fibers is seen emanating from the injection site, with some commissural fibers (CC) continuing medially towards the opposite hemisphere through the corpus callosum. B, the cord segregates into the commissural fibers and the subcortical bundle (SB) that descends into the anterior limb of the internal capsule (ICa). C, caudal to the injection site the subcortical bundle continues its course in the posterior limb of the internal capsule (ICp). Note terminations in the putamen (Put). Mag = .5, bar = 5 mm (adapted from Schmahmann and Pandya, 2006). See list of abbreviations.

The importance of the striatal and subcortical fiber tracts is reflected in clinical syndromes resulting from lesions of the basal ganglia, thalamus and cerebellum. Here we review these syndromes in conjunction with the relevant details of the anatomical connections. This analysis illustrates the close relationship between the known connectional neuroanatomy in non-human primates, and the behavioral manifestations of disruption of the distributed neural systems in humans.

2. Basal ganglia

2.1. Clinical features

Subcortical dementia as a clinical entity was first recognized in progressive supranuclear palsy and Huntington’s disease (Albert et al., 1974), characterized by slowness of mental processing, forgetfulness, apathy, and depression. This notion was later expanded when it became apparent that focal subcortical lesions play a role in arousal, attention, mood, motivation, language, memory, abstraction, and visuospatial skills (Cummings and Benson, 1984). Patients with Parkinson’s disease experience apathy with diminished emotional responsiveness (Dujardin et al., 2007), and cognitive decline including impairment of concentration (bradyphrenia, Naville, 1922), loss of mental and behavioral flexibility with impaired strategic planning, sequential organization, constructional praxis, verbal fluency, working memory, attentional set shifting, initial encoding of information, procedural learning and spontaneous recall, and aspects of language that rely on procedural memory (Growdon and Corkin, 1987; Owen et al., 1992; Cronin-Golomb et al., 1994; Smith and McDowall, 2006; Muslimovic et al., 2007; Williams-Gray et al., 2007). These deficits are thought to reflect damage to fronto-striatal interactions, and interference with the role of basal ganglia in the cognitive processes that lead to habit formation (Barnes et al., 2005) and goal-directed behaviors (Delgado, 2007).

Three broad categories of behavioral–cognitive syndromes are now recognized following damage to the basal ganglia. Cognitive deficits occur with lesions of the rostral head of the caudate nucleus, including impaired working memory, strategy formation, and cognitive flexibility. Focal lesions are associated with impairments true to hemisphere – hemi-neglect and visual–spatial disorientation following right caudate lesions, aphasia after left caudate stroke (Caplan et al., 1990; Kumral et al., 1999). Lesions of the ventral striatum produce disinhibited, irritable and labile behaviors, and are implicated in the neurobiology of addiction (Peoples et al., 2007) and obsessive compulsive disorder (Remijnse et al., 2006). Lesions of the putamen lead to extrapyramidal motor syndromes and akinesia, and apathy and unconcern can follow damage to the dorsolateral striatum (Levy and Dubois, 2006).

2.2. Connectional neuroanatomy of the basal ganglia

Tract tracing studies show that there are multiple parallel loops, or circuits, in the corticostriatal system, each of which comprises a parent cerebral cortical area (motor, association, or limbic cortex) that projects in a topographically arranged manner to nuclei of the basal ganglia, which project in turn via thalamus back to the cortical region of origin (Nauta and Domesick, 1984; Alexander et al., 1986; Selemon and Gold-man-Rakic, 1990; Mega and Cummings, 1994). Each of these segregated loops supports distinct domains of behavior. Sensorimotor and parietal intramodality sensory association cortices project predominantly to the dorsal and mid-sectors of the putamen. Association areas in prefrontal, posterior parietal, and superior temporal polymodal cortices project preferentially to the caudate nucleus. Orbital and medial prefrontal cortices and the cingulate gyrus project to the ventral striatum, and the rostral and inferior temporal and parahippocampal cortices project to the ventral striatum and ventral part of the putamen (Yeterian and Pandya, 1993, 1994, 1998; Schmahmann and Pandya, 2006). All these projections are arranged with a high degree of topographic specificity (Fig. 4).

Fig. 4.

Diagrams illustrating the principle of topographic arrangement of cerebral cortical projections to the striatum from prefrontal cortex (A), posterior parietal cortex (B), and motor cortices (C). Injections of isotope tracer into selected regions of the cerebral cortex are shown at left. A representative coronal image from the striatum is shown on the right, and plane of section of the image is depicted on the cerebral hemisphere image at left. Color-coding denotes the isotope injection sites in the hemispheres, the terminations in the striatum, and the fibers within the Muratoff bundle (MB), external capsule (EC), and the striatal bundle (StB) coursing between the MB and EC. Some association and projection fibers as well as claustrum and cortex terminations are also visible in these composite summary images, but not labeled. A, Prefrontal cortex – medial, lateral, and basal surfaces of the cerebral hemisphere: blue, medial prefrontal convexity area 32; green, area 46d above the mid-portion of the principal sulcus; orange, area 46v below the mid-portion of the principal sulcus, involving sulcal and gyral cortices; red, orbital frontal cortex area 47/12, encroaching posteriorly on the insular proisocortex. B. Posterior parietal cortex – medial, dorsal, and lateral surfaces of the cerebral hemisphere; black, medial parietal convexity area PGm, encroaching upon area PEc; purple, medial part of area PEc at the junction of area PE; blue, lateral part of area PEc at the junction of area PE; green, area PG and area Opt; orange, rostral inferior parietal lobule, area PF; red, midregion of the parietal operculum. C. Motor cortices – medial and lateral surfaces of the cerebral hemisphere; black, frontal operculum in the precentral aspects of areas 1 and 2; orange, ventral area 4, face motor representation; green, area 4 behind the arcuate spur, hand motor representation; blue, dorsal precentral gyrus area 4, trunk motor representation; purple, dorsal area 4, foot motor representation; red, medial part of the superior frontal gyrus, rostral area MII, supplementary motor face representation (adapted from Schmahmann and Pandya, 2006).

The major behavioral–cognitive syndromes that arise following basal ganglia lesions thus reflect the anatomic connections with the cerebral cortex. Deficits in executive function and spatial cognition from lesions of the head of the caudate nucleus (Caplan et al., 1990) reflect the connections with the dorsolateral prefrontal cortex (Yeterian and Pandya, 1994; Schmahmann and Pandya, 2006) concerned with attention and executive functions (Stuss and Benson, 1984), and the posterior parietal cortex (Yeterian and Pandya, 1993; Schmahmann and Pandya, 2006) concerned with personal and extra-personal space (Mesulam, 1981). The precisely arranged topography of the associative projections to the striatum, referred to above, is likely to account for the observation that different regions of striatum engage in different aspects of cognition. Limbic behaviors result from lesions of the ventral striatum reflecting connections with the orbital and medial prefrontal cortices concerned with drive, motivation, emotional aspects of performance, inhibition of inappropriate responses, and reward-guided behaviors (Everitt et al., 1999; Robbins and Everitt, 2002; Levy and Dubois, 2006; Barbas, 2007; Schoenbaum et al., 2007). Whereas apathy most likely reflects involvement of the limbic ventral striato-pallidal system, it can also occur following damage to dorsolateral striatum (Bhatia and Marsden, 1994; Hama et al., 2007), and may reflect connections with the medial prefrontal and anterior cingulate regions. Extrapyramidal motor syndromes, including dystonia and hemiparkinsonism, result from lesions of the dorsal and mid-regions of the putamen that receive afferents from motor cortices (Bhatia and Marsden, 1994). Hypophonic dysarthria is common in Parkinson’s disease, possibly also reflecting involvement of the putamen.

Motor and behavioral consequences of pallidotomy are also determined by lesion location. Rostral and dorsomedial GPi lesions (linked to prefrontal cortical areas 9 and 46) produce impaired semantic fluency, mathematical ability, and memory under conditions of proactive interference. Lesions of posterior and ventrolateral regions (linked to motor cortical areas) have a beneficial impact on bradykinesia, but no influence on cognitive performance (Lombardi et al., 2000). Left-sided lesions produce deficits in verbal fluency and verbal encoding (Trépanier et al., 1998). Bilateral pallidotomy, while generally effective at reducing the disabling features of Parkinson’s disease (Scott et al., 1998), can result in prominent behavioral changes, disinhibition, reckless and socially inappropriate behaviors, apathy, poor judgment and lack of insight (Ghika et al., 1999).

3. Thalamus

3.1. Clinical features

In the first detailed account of the behavioral consequences of thalamic hemorrhage, Fisher (1959) described neglect (“modified anosognosia and hemiasomatognosia”), global “dysphasia”, confusion and visual hallucinations. Accounts followed of thalamic dementia from prion diseases (Martin, 1997), and behavioral changes in patients with thalamic tumors (Ziegler et al., 1977; Nass et al., 2000), but these lesions are seldom confined to thalamus. There are four thalamic vascular syndromes that illustrate the behavioral roles of the thalamic nuclei (Schmahmann, 2003) (Table 2; Fig. 5).

Table 2.

Thalamic arterial supply and principal clinical features of focal infarction

Thalamic blood vessel	Nuclei irrigated	Clinical features reported
Tuberothalamic artery (arises from middle third Of P. Comm)	Reticular, Intralaminar, VA, rostral VL, ventral pole of MD, anterior nuclei – AD, AM, AV, ventral internal medullary lamina, ventral amygdalofugal pathway, mamillothalamic tract	Confusion, Memory, Emotion, Behavior Fluctuating arousal and orientation Impaired learning, memory, autobiographical memory Superimposition of temporally unrelated information Personality changes, apathy, abulia Executive failure, perseveration True to hemisphere – language if VL involved on left; hemispatial neglect if right sided Emotional facial, acalculia, apraxia
Paramedian artery (arises from P1)	MD, intralaminar – CM, Pf, CL, posteromedial VL, ventromedial pulvinar, paraventricular, LD, dorsal internal medullary lamina	Confusion, Memory, Language, Behavior Decreased arousal (coma vigil if bilateral) Impaired learning and memory, confabulation, temporal disorientation, poor autobiographical memory Aphasia if left sided, spatial deficits if right sided Altered social skills and personality, including apathy, aggression, agitation
Inferolateral artery (arises from P2)
Principal inferolateral Branch	Ventroposterior complex – VPM, VPL, VPI	Variable elements of a triad – hemisensory loss, hemiataxia, and hemiparesis Sensory loss (variable extent, all modalities)
	Ventral lateral nucleus, ventral (motor) part	Hemiataxia, hemiparesis, dystonia, tremor Post-lesion pain syndrome (Dejerine-Roussy) – right hemisphere predominant
Medial branch   Inferolateral pulvinar branches	Medial geniculate Rostral and lateral pulvinar, LD nucleus	?Auditory consequences ?Behavioral
Posterior choroidal artery (arises from P2)
Lateral branches   Medial branches	LGN, LD, LP, inferolateral parts of pulvinar MGN, posterior parts of CM and CL, pulvinar	Visual field loss (hemianopsia, quadrantanopsia) Variable sensory loss, weakness, aphasia, memory Impairment

Fig. 5.

Thalamic vascular supply. Schematic diagram of the lateral (A) and dorsal (B) views of the four major thalamic arteries, and the nuclei they irrigate, according to Bogousslavsky et al. (1988). 1 = Carotid artery, 2 = basilar artery, 3 = P1 region of the posterior cerebral artery (mesencephalic artery), 4 = posterior cerebral artery, 5 = posterior communicating artery, 6 = tuberothalamic artery, 7 = paramedian artery, 8 = inferolateral artery, 9 = posterior choroidal artery. P = pulvinar. The illustrations and abbreviations in (C) and (D) from De Freitas and Bogousslavsky (2002) are an adapted version of the conclusions of Von Cramon et al. (1985) regarding the patterns of irrigation by the thalamic arterial supply to the thalamic nuclei. (Note: posterior choroidal artery irrigation of the anterior thalamus is disputed – see Schmahmann, 2003).

3.1.1. Tuberothalamic artery infarction (Fig. 6A, B)

Fig. 6.

A, Diffusion weighted image (DWI) of left tuberothalamic artery territory infarction (right of diagram). B, DWI showing infarction bilaterally in the territory of the polar branch of the tuberothalamic artery, likely representing an example of the paramedian artery irrigating both the paramedian and tuberothalamic “territories”. C, right paramedian artery territory infarction, seen on T2 weighted MRI. D, acute infarction in the left inferolateral artery territory on DWI.

These patients demonstrate fluctuating levels of consciousness, disorientation in time and place, and personality changes include euphoria, lack of insight, apathy, lack of spontaneity, and emotional unconcern (Graff-Radford et al., 1984, 1985; Bogousslavsky et al., 1986, 1988; Lisovoski et al., 1993). New learning and verbal and visual memory are impaired. Amnesia is greater following left thalamic lesions, along with anomia, impaired comprehension, fluent and meaningless discourse with semantic and phonemic paraphasic errors, neologisms and perseveration. In contrast, repetition and reading aloud are preserved. Acalculia, and buccofacial and limb apraxia may occur (Warren et al., 2000). Right thalamic lesions produce the amnestic syndrome along with impairments in visual processing and visual memory, and hemispatial neglect. Lesions on either side produce emotional central facial paralysis (good facial movement with volition, but facial asymmetry during emotional display), and constructional apraxia. Motor findings are mild, and sensory disturbances are rare.

Stroke confined to the anterior/polar branch of the tuberothalamic artery affects memory and personality. Autobiographic memory and newly acquired information are disorganized with respect to temporal order, with patients displaying superimposition of temporally unrelated information (palipsychism). Apathy, inattention, disorientation, impaired sequencing and perseveration are prominent. Dysarthria, hypophonia, anomia and decreased verbal fluency are noted, but comprehension, writing, reading, and repetition are preserved (Clarke et al., 1994; Ghika-Schmid and Bogousslavsky, 2000).

3.1.2. Paramedian artery infarction (Fig. 6C)

Disturbances of arousal and memory, confusion, agitation, aggression, apathy and perseveration are common (Castaigne et al., 1981; Graff-Radford et al., 1984, 1985; Bogousslavsky et al., 1988). Left thalamic strokes produce adynamic aphasia (Guberman and Stuss, 1983) with reduced verbal fluency, preserved syntax, occasional paraphasias and normal repetition. Right paramedian strokes impair visual– spatial functions. Bilateral infarction produces disorientation, confusion, hypersomnolence, and akinetic mutism (awake unresponsiveness) (Castaigne et al., 1981; Graff-Radford et al., 1984, 1985; Reilly et al., 1992) as well as severe anterograde and retrograde memory deficits, apathy, inappropriate social behaviors, and a reported absence of spontaneous thoughts or mental activities (see Guberman and Stuss, 1983; Bogousslavsky et al., 1991; Engelborghs et al., 2000). Distortion of personally relevant autobiographical memory with relative sparing of knowledge of famous people and public events has been observed – a thematic retrieval memory disorder (Hodges and McCarthy, 1993). Disorientation in time (chronotaraxis, Spiegel et al., 1956), apraxia and dysgraphia are also reported (Castaigne et al., 1981; Graff-Radford et al., 1990).

3.1.3. Inferolateral artery infarction (Fig. 6D)

The thalamic syndrome of Dejerine and Roussy (1906) includes sensory loss, hemiparesis and post-lesion pain, particularly following right-sided lesions (Garcin and Lapresle, 1954; Lapresle and Haguenau, 1973; Fisher, 1978; Bogousslavsky et al., 1988; Caplan et al., 1988; Nasreddine and Saver, 1997). Sensory loss may be the sole clinical manifestation, involve all modalities, or impair spinothalamic sensation (temperature, pinprick) without loss of posterior column sense (position, vibration). These strokes produce a flexed and pronated “thalamic hand” (Foix and Hillemand, 1925) with the thumb buried beneath the other fingers; and ataxia and hemiparesis in the same extremities (Dejerine and Roussy, 1906; Foix and Hillemand, 1925; Garcin, 1955; Caplan et al., 1988; Gutrecht et al., 1992). Myoclonic dystonia, dystonic postures with slow, pseudo-athetoid movements, and postural and action tremor may also be observed (Lehéricy et al., 2001). Cognitive and psychiatric presentations are notably absent.

3.1.4. Posterior choroidal artery infarction

These strokes produce complex visual field deficits reflecting involvement of the lateral geniculate nucleus, hemisensory loss, transcortical aphasia, and memory deficits (Bogousslavsky et al., 1988), and more rarely, a delayed complex hyperkinetic motor syndrome with ataxia, rubral tremor, dystonia, myoclonus and chorea (Neau and Bogousslavsky, 1996). Spatial neglect has been reported (Karnath et al., 2002).

3.2. Connectional neuroanatomy of the thalamus

These clinical syndromes may be understood by considering the functional properties of the thalamic nuclei, as determined by tract tracing investigations in monkeys, and physiological and clinical studies in patients. We group the thalamic nuclei into six functional classes – the reticular nucleus engaged in arousal; intralaminar nuclei subserving attention and nociception; limbic nuclei concerned with mood and motivation; specific sensory nuclei; effector nuclei concerned with movement and aspects of language; and associative nuclei participating in high level cognition (Schmahmann, 2003; Fig. 7; Table 3).

Fig. 7.

Diagram illustrating the nuclei of the human thalamus. Horizontal sections are seen above, from ventral to dorsal. Coronal sections below proceed from rostral to caudal. The revised nomenclature correlates with terminology used in the monkey. An earlier nomenclature is shown in parenthesis. See list of abbreviations (from Jones, 1997).

Table 3.

Behavioral roles of thalamic nuclei (see list of abbreviations)

Major functional grouping	Thalamic nuclei	Putative functional attributes
Reticular	Reticular	Arousal, rhythmicity, role in epileptogenesis
Intralaminar	CM, Pf, CL, Pcn, midline (reunions, paraventricular, rhomboid)	Arousal, attention, motivation, affective components of pain
Limbic	Anterior nuclear group (AD, AM, AV) Lateral dorsal nucleus Other – MDmc, medial pulvinar, ventral anterior	Learning, memory, emotional experience and expression, drive, motivation
Specific sensory	Medial geniculate	Auditory
	Lateral geniculate	Visual
	Ventroposterior   Lateral (VPL)   Medial (VPM)   Medial, parvicellular (VPMpc)   Inferior (VPI)	Somatosensory body and limbs Somatosensory head and neck Gustatory Vestibular
Effector	Ventral anterior   Reticulata recipient   Pallidal recipient   Ventral medial	Complex behaviors Motor programming Motor
	Ventral lateral   Ventral part   Dorsal part	Motor Language (dominant hemisphere)
Associative	Lateral posterior	High order somatosensory and visuospatial integration – spatial cognition
	Medial dorsal   Medial, magnocellular (MDmc)   Intermediate, parvicellular (MDpc)   Lateral, multiform (MDmf)	Drive, motivation, inhibition, emotion Executive functions, working memory Attention, horizontal gaze
	Pulvinar   Medial   Lateral   Inferior   Anterior (pulvinar oralis)	Supramodal, high level association region across multiple domains Somatosensory, visual association Visual association Intramodality somatosensory association, pain appreciation

4. Reticular thalamic nucleus

This nuclear shell surrounds thalamus and conveys afferents from cerebral cortex. It contributes to synchrony and rhythms of thalamic neuronal activity, and is relevant in the pathophysiology of epilepsy (Huguenard and Prince, 1997), and the neural substrates of consciousness (Llinas and Ribary, 2001).

5. Intralaminar thalamic nuclei

The paracentral (Pcn), central lateral (CL), centromedian (CM), parafascicular (Pf) and midline nuclei such as paraventricular, rhomboid and reunions play a role in autonomic drive. They receive afferents from brainstem, spinal cord, and cerebellum, and have reciprocal connections with cerebral hemispheres (Brodal, 1981; Jones, 1985). The CM/Pf nuclei are also linked with the basal ganglia in tightly connected functional circuits. A sensorimotor circuit links putamen with CM through the ventrolateral part of the internal segment of the globus pallidus (GPi); a limbic circuit links the ventral striatum with the Pf through the rostromedial GPi; and cognitive circuits link the caudate nucleus with Pf through the dorsal GPi and through the substantia nigra pars reticulata (Sidibe et al., 2002). Midline nuclei receiving input from the periaqueductal gray and spinothalamic tract are involved in processing the motivational–affective components of pain (Bentivoglio et al., 1993; Lenz and Dougherty, 1997; Willis, 1997).

6. Limbic thalamic nuclei

The functions of the anterior nuclear group – ventral, medial, and dorsal (AV, AM and AD nuclei), and the lateral dorsal (LD) nucleus, reflect their reciprocal anatomical connections with limbic structures in the cingulate gyrus, hippocampus, parahippocampal formation, entorhinal cortex, retrosplenial cortex, orbitofrontal and medial prefrontal cortices, and with subcortical structures including the mamillary bodies and amygdala (Yakovlev et al., 1960; Locke et al., 1961; Yakovlev and Locke, 1961; Yeterian and Pandya, 1988). The magnocellular part of the medial dorsal nucleus (MDmc), parts of the medial pulvinar, and parts of the VA nucleus are also reciprocally interconnected with the cingulate gyrus and other components of the limbic system (Goldman-Rakic and Porrino, 1985; Vogt et al., 1987; Siwek and Pandya, 1991), and thus may also be considered limbic. Like their cortical and subcortical counterparts (Mesulam, 1988; Devinsky and Luciano, 1997), limbic thalamic nuclei are likely to be relevant for learning and memory, emotional experience and expression, drive and motivation. The tuberothalamic artery irrigates these nuclei as well as the mammillothalamic and ventral amygdalo-fugal tracts that link the anterior thalamic nuclei with the limbic regions (Graff-Radford et al., 1990), accounting for the profound amnestic and limbic deficits resulting from tuberothalamic stroke.

7. Specific sensory thalamic nuclei

The specific sensory nuclei include the medial geniculate nucleus (MGN), lateral geniculate nucleus (LGN), and ventroposterior nuclei (lateral, medial, and inferior – VPL, VPM VPI).

Medial geniculate nucleus connections with primary and association auditory cortices infer a role in higher level auditory processing, as well as in elementary audition (Mesulam and Pandya, 1973; Pandya et al., 1994; Hackett et al., 1998).

The lateral geniculate projects to primary and secondary visual cortices (Kennedy and Bullier, 1985). It also receives projections back from visual areas (Shatz and Rakic, 1981), indicating that higher order processing can influence visual perception at an early stage.

The VPL and VPM nuclei are reciprocally interconnected with primary somatosensory cortices; VPL serves body and limbs, VPM head and neck (Jones and Powell, 1970a, 1970b). Gustatory function is subserved by the parvicellular division of VPM (Pritchard et al., 1986). The somatotopy of these nuclei is precise, and lacunar infarcts of the inferolateral artery produce focal sensory deficits. The VPI nucleus is linked with the rostral inferior parietal lobule and the second somatosensory area in the parietal operculum (SII, Yeterian and Pandya, 1985; Schmahmann and Pandya, 1990) and with the frontal operculum engaged in vestibular functions (Deecke et al., 1977).

Spinothalamic and trigeminothalamic inputs, and topographically organized wide dynamic neurons and nociceptive specific neurons in the ventroposterior nuclei facilitate the role of these nuclei in the specific component of the pain system (Willis, 1997). Disruption of SII cortex connections with these nuclei has been postulated to cause the parietal pseudothalamic pain syndrome (Schmahmann and Leifer, 1992).

8. Effector thalamic nuclei

Motor nuclei include the ventral anterior (VA), ventromedial (VM), and ventral lateral (VL) nuclei. Subregions within VA receive afferents from the internal globus pallidus (Ilinsky and Kultas-Ilinsky, 1987), are linked with premotor cortices (Jones, 1997), and may be responsible for dystonia in rostral thalamic lesions. Neurons in VA receiving afferents from the substantia nigra pars reticulata (Jones, 1985; Francois et al., 2002) are linked with premotor, supplementary motor (Schell and Strick, 1984), prefrontal (Goldman-Rakic and Porrino, 1985), caudal parts of the posterior parietal (Schmahmann and Pandya, 1990), and rostral cingulate cortices (Vogt et al., 1987), and may account for complex behavioral syndromes following lesions of the anterior thalamus.

The ventral sector of the posterior part of VL is linked with the motor cortex (Strick, 1976), and causes ataxia and mild motor weakness following thalamic stroke (Murthy, 1988; Gutrecht et al., 1992; Solomon et al., 1994). The dorsal sector is linked with the posterior parietal (Schmahmann and Pandya, 1990), prefrontal (Kievit and Kuypers, 1977; Künzle and Akert, 1977; Middleton and Strick, 1994), and superior temporal cortices (Yeterian and Pandya, 1989), and has a role in articulation and language. Perseveration results from electrical stimulation of left medial VL; misnaming and omissions with stimulation of left posterior VL. The dorsal sector of VL is also engaged in the encoding and retrieval of verbal (left) and nonverbal information (right) (Ojemann et al., 1968; Ojemann and Ward, 1971; Johnson and Ojemann, 2000; Hugdahl and Wester, 2000).

9. Associative thalamic nuclei

The lateral posterior, medial dorsal, and pulvinar nuclei are interconnected with cerebral association areas, and have no peripheral afferents or links with primary sensorimotor cortices.

The lateral posterior (LP) nucleus is reciprocally linked with the posterior parietal (Weber and Yin, 1984; Yeterian and Pandya, 1985; Schmahmann and Pandya, 1990), medial and dorsolateral extrastriate (Yeterian and Pandya, 1997), and posterior cingulate and medial parahippocampal cortices (Yeterian and Pandya, 1988). It is able to integrate intramodal and multimodal associative somatosensory and visual information, and is likely engaged in spatial functions, goal directed reaching (Acuña et al., 1990), and possibly in conceptual and analytical thinking.

The medial dorsal (MD) nucleus has reciprocal connections with the prefrontal cortex (Tobias, 1975; Goldman-Rakic and Porrino, 1985; Giguere and Goldman-Rakic, 1988; Barbas et al., 1991; Siwek and Pandya, 1991). The medial part (magnocellular MDmc) is linked with paralimbic regions – medial and orbital prefrontal cortices, amygdala, basal forebrain, and olfactory and entorhinal cortices (Russchen et al., 1987; Graff-Radford et al., 1990). Apathy, abulia, disinhibition, and failure to inhibit inappropriate behaviors are likely to result from MDmc lesions, and possibly memory (Victor et al., 1971) and language deficits (Bogousslavsky et al., 1988) as well. The intermediate part of MD (the parvicellular MDpc) is linked with dorsolateral and dorsomedial prefrontal cortices, areas 9 and 46, possibly accounting for poor working memory and perseveration resulting from MD lesions. The laterally placed multiformis part (MDmf) is linked with the area 8 in the arcuate concavity, and lesions produce impairments of horizontal gaze and attention.

Different subregions within medial pulvinar (PM) are topographically linked with prefrontal (Asanuma et al., 1985; Yeterian and Pandya, 1988; Romanski et al., 1997), posterior parietal (Asanuma et al., 1985; Yeterian and Pandya, 1985; Schmahmann and Pandya, 1990), and auditory related (Pandya et al., 1994) and multimodal superior temporal cortices (Yeterian and Pandya, 1989), and with the cingulate, parahippocampal (Yeterian and Pandya, 1988) and insula cortices (Mufson and Mesulam, 1984). Aphasia(Ojemann et al., 1968), spatial neglect (Karnath et al., 2002) and psychosis (Guard et al., 1986) may results from PM lesions. The lateral pulvinar (PL) is linked with posterior parietal (Asanuma et al., 1985; Schmahmann and Pandya, 1990), superior temporal (Yeterian and Pandya, 1989), and medial and dorsolateral extrastriate cortices (Yeterian and Pandya, 1997), and the superior colliculus (Robinson and Cowie, 1997). It is engaged in the integration of somatosensory and visual information. Inferior pulvinar (PI) is linked with temporal lobe areas concerned with visual feature discrimination, and with ventrolateral and ventromedial extrastriate areas concerned with visual motion (Cusick et al., 1993; Yeterian and Pandya, 1997). It also receives input from retinal ganglion cells (Cowey et al., 1994) and visual neurons of the superior colliculus (Robinson and Cowie, 1997). The anterior pulvinar (pulvinar oralis, PO) is interconnected with intramodality somatosensory association cortices in the rostral part of the inferior parietal region, and with the second somatosensory area (SII, Asanuma et al., 1985; Yeterian and Pandya, 1985; Acuña et al., 1990; Schmahmann and Pandya, 1990). The PO nucleus may be important in the appreciation of pain, as are the suprageniculate, limitans, and posterior nuclei (Jones, 1985).

10. Thalamic connectional topography

Sensorimotor, effector, limbic, and associative regions of cerebral cortex are therefore linked with distinctly different sets of thalamic nuclei. Thalamic projections to the posterior parietal lobe exemplify this concept (Schmahmann and Pandya, 1990). Connections become progressively elaborated as one moves from rostral to caudal within both the superior and the inferior parietal lobules. Rostral areas concerned with intramodality somatosensory processing are related to modality-specific thalamic nuclei, whereas caudal regions, concerned with complex functions, derive their input from multimodal and limbic nuclei (Fig. 8). This rostral-caudal cortical topography is represented within the lateral posterior and pulvinar oralis nuclei that project to both superior and inferior parietal lobules. Rostral parietal subdivisions receive projections from ventral regions within these thalamic nuclei, caudal parietal afferents arise from the dorsal parts of these nuclei, and the intervening cortical levels receive projections from intermediate positions within the nuclei. A similar topographic arrangement is present also in the medial pulvinar projections to the inferior parietal lobule (Fig. 9).

Fig. 8.

Diagrammatic representation of projections from thalamus (retrogradely labeled neurons identified as black dots) to primary somatosensory cortex S1, area PF, area PG, and area PG-–Opt of the parietal lobe in rhesus monkey following cortical injections (blackened areas) of wheat germ agglutinated horseradish peroxidase. Representative rostral to caudal levels of thalamus are shown. (Thalamic nomenclature according to Olszewski, 1952. Parietal lobe nomenclature according to Pandya and Seltzer, 1982.) The table above summarizes the areas injected with tracer, the principal thalamic nuclei demonstrating retrogradely labeled neurons, and the putative functional attributes of the cortical areas studied. See list of abbreviations (derived from Schmahmann and Pandya, 1990).

Fig. 9.

Diagrammatic representation of the projections from the medial pulvinar nucleus of thalamus (PM) to the inferior parietal lobule in rhesus monkey. Fluorescent retrograde tracers were placed in area PF (blue), area PG (red), and area PG–Opt (green) in a rhesus monkey, and the resulting retrogradely labeled neurons were identified in the PM nucleus. See list of abbreviations (adapted from Schmahmann and Pandya, 1990).

Damage to thalamic fiber pathways may have clinical consequences. The mamillothalamic tract (MMT) connects the anterior thalamic nuclei with the mamillary body, which is linked with hippocampus and entorhinal cortex (Crosby et al., 1962; Brodal, 1981). The MMT and fornix bind the anterior thalamic nuclei into the neural system for learning and memory. The ventral amygdalofugal pathway links the amygdala with the medial part of MD, and damage contributes to amnesia and emotional dysregulation (Graff-Radford et al., 1990). Lesions of the superior, medial and inferior, and lateral thalamic peduncles, and of the anterior limb of the internal capsule (conveying prefrontal and anterior cingulate interactions; Schmahmann and Pandya, 2006) can produce corticothalamic disconnection and complex behavioral syndromes (Schmahmann, 1984; Tatemichi et al., 1992; Chukwudelunzu et al., 2001). Disorders of eye movement result from medial thalamic lesions destroying descending tracts from motor and premotor cortices to the midbrain nuclei of Darkschewitz and the interstitial nucleus of Cajal (up and down gaze), and the rostral nucleus of the medial longitudinal fasciculus in the tectum (downgaze) (Fisher, 1959; Guberman and Stuss, 1983; Leigh and Zee, 1983; Bogousslavsky et al., 1988).

11. Cerebellum

11.1. Clinical features

The cerebellum is subcortical only in the sense that it is distinct from cerebral cortex. The traditional view that cerebellar function is confined to the coordination of voluntary motor activity has evolved in recent years (Schmahmann, 1997). Evidence from patients has made it plain that cerebellar pathology is related to intellectual and emotional deficits in addition to motor incoordination. The wider role of the cerebellum in nervous system function has far-reaching implications for understanding the neural substrates of higher order behavior and neuropsychiatric disorders. There appears to be a double dissociation in the organization of motor and non-motor functions in the cerebellum, a conclusion derived from clinical neurology as well as anatomy, physiology and functional imaging studies.

Clinical examination of patients with cerebellar stroke demonstrates that the cerebellar motor syndrome of gait ataxia, appendicular dysmetria, dysarthric speech and oculomotor abnormalities results from lesions that affect the anterior lobe of the cerebellum, notably lobules I through V, with lobule VI (Fig. 10A) likely engaged in motor control in a manner possibly equivalent to premotor regions of the cerebral cortex (Schmahmann, 2007). There is a second sensorimotor area identified by physiology (Snider, 1950; Snider and Eldred, 1952) and functional imaging (Grodd et al., 2001), located in lobule VIII at the medial part of the posterior lobe. Oculomotor abnormalities, along with prominent vestibular symptoms (vertigo, nausea, emesis), arise from lesions that involve lobules IX and X of the posterior and flocculonodular lobes (Duncan et al., 1975; Lee et al., 2006). Remarkably, large lesions in the major expansion of the cerebellar hemisphere, i.e., lobules crus I and II and lobule VIIB in the posterior lobe (Fig. 10B), do not result in the cerebellar motor syndrome. Indeed, when these patients are examined a few days after stroke, and the vestibular symptoms have subsided, it can be difficult to detect any sign of a motor disorder, a straight-forward clinical finding that defies two centuries of dogma about the cerebellar role being confined exclusively to motor coordination (Schmahmann, 2007).

Fig. 10.

Diffusion weighted MRI of cerebellar infarction. A, Reference diagram in the horizontal plane of the major cerebellar lobular groupings (top left: superior, to bottom right: inferior). The anterior lobe (lobules I through V) is shaded black; lobule VI of the posterior lobe is shaded gray; lobules VII through IX of the posterior lobe, and lobule X (flocculonodular lobe) are not shaded. B, stroke in the territory of the superior cerebellar artery, involving the anterior lobe (cerebellar lobules I–V) and part of the posterior lobe. This patient had a cerebellar motor syndrome, scoring 20 of possible 120 points on the Modified International Cooperative Ataxia Rating Scale (MICARS, Schmahmann, 2007; Schmahmann et al., 2007a; Trouillas et al., 1997). C, stoke in the posterior inferior cerebellar artery territory, sparing the anterior lobe. The patient was motorically normal, with a MICARS score of 1.

The other side of the double dissociation is that the cerebellar cognitive affective syndrome (CCAS, Schmahmann and Sherman, 1998) occurs following lesions of the cerebellar posterior lobe, but not the anterior lobe (Schmahmann and Sherman, 1998; Exner et al., 2004). The CCAS is characterized by deficits in executive function, visual spatial performance, linguistic processing and affective dysregulation. Executive impairments include deficits in working memory, motor or ideational set shifting, and perseveration. Verbal fluency may be impaired to the point of telegraphic speech or mutism. Visuospatial disintegration impairs attempts to draw or copy a diagram, conceptualization of figures can be disorganized, and some patients display simultanagnosia. Anomia, agrammatic speech and abnormal syntactic structure are observed, with abnormal prosody and occasionally high pitched, hypophonic whining. Abnormal modulation of behavior and personality is notable with posterior lobe lesions that involve the vermis and fastigial nucleus. This manifests as flattening of affect, alternating or coexistent with disinhibited behaviors such as over-familiarity, flamboyant and impulsive actions, and humorous but inappropriate and flippant comments. Regressive, childlike behaviors and obsessive-compulsive traits can be observed. Autonomic changes have been noted following lesions of the fastigial nucleus and vermis, manifesting as bradycardia and syncope, or tachycardia in the setting of acquired panic disorder (Haines et al., 1997; Schmahmann and Sherman, 1998; Schmahmann et al., 2007c). The principal features and clinical relevance of the CCAS have been replicated in adults with stroke (Malm et al., 1998; Leggio et al., 2000; Neau et al., 2000), in children who have undergone excision of cerebellar tumors (Levisohn et al., 2000; Riva and Giorgi, 2000; Scott et al., 2001), and in other acquired and developmental disorders of the cerebellum (Ciesielski et al., 1999; Steinlin et al., 1999; Allin et al., 2001; Chheda et al., 2002; van Harskamp et al., 2005; Limperopoulos et al., 2006; Tavano et al., 2007). Deficits in emotional expression in patients with strokes and neurodegenerative lesions have also been linked to the involvement of the cerebellar system (Parvizi et al., 2001, 2007).

The posterior fossa syndrome represents a particularly acute form of the CCAS (Wisoff and Epstein, 1984; Pollack, 1997; Levisohn et al., 2000; Riva and Giorgi, 2000). Within 48 h following surgical resection of midline tumors of the cerebellum, children develop mutism, buccal and lingual apraxia, apathy and poverty of spontaneous movement. Emotional lability is marked by rapid fluctuation between irritability and agitation to giggling and easy distractibility. The range of behavioral impairments in the setting of cerebellar lesions over and above the intellectual deficits is remarkably wide. The neuropsychiatry of the cerebellum may be conceptualized as falling into five major behavioral domains– attentional control, emotional control, social skill set, autism spectrum disorders, and psychosis spectrum disorders (Schmahmann et al., 2007c) (Table 4).

Table 4.

Neuropsychiatric manifestations in cerebellar disorders

	Positive (exaggerated) symptoms	Negative (diminished) symptoms
Attentional Control	Inattentiveness Distractibility Hyperactivity Compulsive and ritualistic behaviors	Ruminativeness Perseveration Difficulty shifting focus of attention Obsessional thoughts
Emotional control	Impulsiveness, disinhibition Lability, unpredictability Incongruous feelings, pathological laughing/crying Anxiety, agitation, panic	Anergy, anhedonia Sadness, hopelessness Dysphoria Depression
Autism spectrum	Stereotypical behaviors Self stimulation behaviors	Avoidant behaviors, tactile defensiveness Easy sensory overload
Psychosis spectrum	Illogical thought Paranoia	Lack of empathy Muted affect, emotional blunting, apathy
Social skill set	Anger, aggression Irritability Overly territorial Oppositional behavior	Passivity, immaturity, childishness Difficulty with social cues and interactions Unawareness of social boundaries Overly gullible and trusting

11.2. Connectional neuroanatomy of the cerebellum

The cerebellar linkage with the cerebral cortex has traditionally been regarded as consisting, in large part, of motor related cortices projecting through basis pontis nuclei to cerebellum, and dentate nucleus of cerebellum sending efferents back through VL thalamus to motor cortex (Crosby et al., 1962; Brodal, 1981). More recent anatomic examination of the cerebrocerebellar system, however, reveals that this system is more intricately organized, and is constructed to subserve the multitude of sensorimotor as well as higher order behavior patterns identified in the clinic, in functional imaging studies, and in experimental observations.

The association and paralimbic regions of the cerebral cortex have topographically organized feedforward projections through the nuclei in the basis pontis into the cerebellum, as well as feedback projections from the cerebellum. Association cortex projections arise from the prefrontal, posterior parietal, superior temporal polymodal regions, and dorsal parastriate cortices. Paralimbic projections arise from the posterior parahippocampal cortex, limbic regions of the cingulate gyrus and the anterior insular cortex involved in autonomic and pain modulation systems (Schmahmann, 1996; Schmahmann and Pandya, 1997a, 1997b). These corticopontine pathways are funneled through the cerebrocerebellar circuit within multiple parallel but partially overlapping loops converging with topographic ordering throughout the pons, whereas the motor corticopontine projections are mostly in the caudal half of the pons (Brodal, 1978; Wiesendanger et al., 1979; Hartmann-von Monakow et al., 1981; Glickstein et al., 1985; Schmahmann and Pandya, 1997b; Schmahmann et al., 2004) (Fig. 11).

Fig. 11.

Color-coded summary diagrams of corticopontine projections from association paralimbic, and motor cortices in rhesus monkey. Left panel: A, medial, B, lateral, and C, orbital surfaces of the cerebral hemisphere coded as follows – prefrontal cortices (purple), posterior parietal (blue), superior temporal (red), posterior parahippocampal and parastriate (orange), and motor and supplementary motor cortices (green). Nine rostral to caudal levels of the ipsilateral basis pontis (I–IX) are depicted, sectioned in the plane shown in the schematic of the brainstem below. Areas in the cerebral hemispheres in yellow and gray do not project to the pons (from Schmahmann, 1996). Right panel: isotope was injected into selected architectonic regions of prefrontal cortex in rhesus monkeys. A, lateral, and B, medial views of the frontal lobe. Color-coded terminations are mostly in the medial part of the pons, and each cortical area has its own unique set of terminations (from Schmahmann and Pandya, 1997a, 1997b).

A precise pattern of organization is probably present also in the pontine projections to the cerebellar cortex, although this still remains to be demonstrated. Physiological and anatomical studies show that the cerebellar anterior lobe and part of lobule VI receive afferents from motor and premotor cortices, whereas the association areas in the prefrontal and posterior parietal cortices are linked predominantly with crus I and II of the posterior lobe (Allen and Tsukahara, 1974; Kelly and Strick, 2003) (Fig. 12). Once conveyed to the cerebellar cortex, these streams of information are acted upon by the cerebellar corticonuclear microcomplexes (Ito, 1984), and then transmitted via the deep cerebellar nuclei to thalamus, on their way back to the cerebral cortex (for a description of the human cerebellar connections see Catani and Thiebaut de Schotten, 2008, this issue).

Fig. 12.

A, lateral view of a cebus monkey brain (top) to show the location of injections of McIntyre-B strain of Herpes simplex virus type 1 in the primary motor cortex, ventral premotor cortex, and areas 9 and 46. The resulting retrogradely labeled neurons in the cerebellar dentate nucleus (below) are indicated by solid dots (from Middleton and Strick, 1997). B, Flattened views of cerebellar cortex of monkey showing areas labeled in anterograde and retrograde fashion following injection of neurotropic virus tracers in the primary motor cortex at left, and prefrontal cortex area 46 at right (adapted from Kelly and Strick, 2003).

Cerebellar projections to thalamus arise from fastigial and interpositus nuclei as well as from the dentate nucleus, and are directed not only to the cerebellar recipient VL, but also to CL, Pcn, the CM Pf complex, and MD that have efferent projections to association cortices. The cerebellar dentate nucleus sends projections through thalamus to different areas of the frontal lobe in the monkey (Middleton and Strick, 1997). The dorsomedial part of the dentate nucleus sends its projections to the motor cortex, whereas the ventrolateral and ventromedial parts of the dentate nucleus are connected with the prefrontal cortex, including area 9/46.

The cerebrocerebellar system thus consists of discretely organized parallel anatomic subsystems that serve as the substrates for differentially organized functional subsystems (or loops) within the framework of distributed neural circuits. This anatomical organization is the substrate for our dysmetria of thought theory (Schmahmann, 1991, 1996, 2000, 2004), and the original notion of the Universal Cerebellar Transform. We have proposed that the cerebellum plays an essential role in automatization and optimizing behavior around a homeostatic baseline according to context; that the cerebellum modulates cognition and emotion in the same way that it coordinates motor control; and that disruption of the neural circuitry linking the cerebellum with the association and paralimbic cerebral regions prevents the cerebellar modulation of functions subserved by the affected subsystems, thereby impairing the regulation of movement, cognition and emotion. This loss of the “cerebellumizing” of behavior leads not only to gait and appendicular ataxia, dysarthria and oculomotor abnormalities when the motor cerebellum is involved, but also to the various aspects of the cerebellar cognitive affective syndrome (Schmahmann and Sherman, 1998) when the cognitive and limbic cerebellar regions are damaged.

12. Discussion

This overview of the neurology and neuropsychiatry of basal ganglia, thalamus and cerebellum makes it clear that “higher cortical functions” are not exclusively the domain of the cerebral cortex. The principle of organization of cerebral cortex, apparent from tract tracing studies, is that each cortical area has association, striatal, and projection fibers linking it in a precise manner to topographically arranged sectors within other cortical regions and subcortical areas. How does the detailed connectional anatomy of each node manifest functionally and clinically, and what is the specific contribution of each of these subcortical nodes to the generation of the observed behaviors? Interspecies hodological principles appear to be supported in that the monkey anatomy agrees in general with the human clinical repertoire, but do the anatomical arrangements identified in monkey hold true for the human brain? In addressing these issues, we present a conceptual approach that, in our view, captures the essential elements of the structure and function of the distributed neural system.

12.1. Evolutionary considerations

Cerebral cortical function may be regarded from the evolutionary perspective as having its origins, both connectionally and functionally, in the subcortical regions. Sensorimotor function, emotion, and rudimentary elements of cognition are governed in lower creatures not by the cerebral cortex, which is either primitive or non-existent, but primarily by subcortical structures – primitive forebears of the basal ganglia, thalamus and cerebellum discussed here, as well as other areas we have not addressed, including the amygdala, hypothalamus and basal forebrain ( Ariëns-Kappers et al., 1936; MacLean, 1972; Karten, 1997). As the cerebral cortex evolved, it became superimposed upon subcortical structures in a systematic and predictable manner (Karten, 1997; Northcutt and Kaas, 1995; McHaffie et al., 2005). Further, according to the mosaic hypothesis, brain evolved by size change in arrays of functionally connected structures (Whiting and Barton, 2003). This has direct bearing on subcortical structures, as exemplified by the cerebrocerebellar system in which the expansion and elaboration of the prefrontal areas occurred in parallel with that of the lateral cerebellar hemispheres (Whiting and Barton, 2003) and the ventral lateral dentate nucleus (Leiner et al., 1986), and the interconnecting fiber bundles in the medial aspect of the cerebral peduncle (Ramnani et al., 2006). The resulting cortico–cortical and cortico–subcortical interactions, linked together in a precise and topographically arranged manner by the connecting white matter tracts, has enabled a more sophisticated repertoire of behaviors. Against this background, we observe that a cortical area may be defined by its pattern of connections; that lesions within subcortical structures can mimic deficits resulting from lesions of cerebral cortical regions with which they are interconnected; and that there is a qualitative difference between the clinical manifestations of lesions in functionally related areas of different cortical and subcortical nodes.

12.2. Architecture determines function – the theory of the universal transform

There is a range of clinical syndromes that results from lesions of different regions within the basal ganglia, thalamus and cerebellum. It appears likely, therefore, that for each node in the distributed neural system, the architecture and connections are reflected in function and clinical manifestations. To account for this, we adopt a reductionist approach derived from an examination of the organizational principles and detailed connectional neuroanatomy. In this view, the architecture and microconnectivity unique to each cortical and subcortical structure facilitates a neural computation, or transform, that is unique to that region. The term “universal” denotes a functional transform at the synaptic/physiological level that is consistent throughout an architectonically defined node.

This universal transform theory was originally introduced with reference to the cerebellum. The recognition that cerebellum is engaged in cognitive processing and emotional regulation as well as motor control, mandated a new model that takes these disparate cerebellar roles into consideration. The early notion that the role of the cerebellum is to modulate neurologic function (Snider, 1950; Dow, 1974; Heath, 1977) is compelling, and we have adopted and amended this as part of a conceptual approach to cerebellar function. We introduced the dysmetria of thought hypothesis (Schmahmann, 1991, 1996, 2000, 2004), which holds that the cerebellum modulates cognitive and emotional processing, as well as motor control. In this view, because cerebellar cortical histology (Ito, 1984; Voogd and Glickstein, 1998), and cerebellar corticonuclear relations (the corticonuclear micro-complex; Ito, 1984) are essentially uniform throughout, the basic work that cerebellum does in the nervous system should be constant as well, hence the Universal Cerebellar Transform (UCT). The UCT is defined as the cerebellar modulation of behavior, serving as an oscillation dampener, maintaining function automatically around a homeostatic baseline, smoothing out performance in all domains. Anatomical subcircuits in the cerebrocerebellar system, discussed above, enable topographically precise interactions between cerebellum and cerebral cortex. These anatomic subcircuits provide the structural basis for putative functional subunits, and facilitate topographic organization of motor and cognitive function in the cerebellum. The computational nature of the UCT remains a matter of intense debate, including notions of oscillation dampening as defined in our dysmetria of thought hypothesis, and various forms of learning (Barto et al., 1999; Leggio et al., 1999; Medina and Mauk, 2000; Molinari et al., 2002; Bloedel, 2004), timing (Ivry, 1997), prediction and preparation (Courchesne and Allen, 1997); sensory data acquisition (Bower, 1997), feedforward internal modeling (Miall et al., 1993; Miall and Reckess, 2002; Ito, 2005; Ramnani, 2006), and others.

According to the dysmetria of thought hypothesis, the UCT is the essential functional contribution that the cerebellum makes to the distributed neural system. By corollary, therefore, there should be a Universal Cerebellar Impairment (UCI). This UCI, the hypothesis holds, is dysmetria, a word derived from the Greek term for disordered timing. When the dysmetria involves the motor domain, the various manifestations of ataxia are evident in extremity movements, eye movements, speech and equilibrium. In contrast, when the dysmetria involves non-motor functions subserved by cerebellum, this results in dysmetria of thought, or cognitive dysmetria, and manifests as the various components of the cerebellar cognitive affective syndrome (Schmahmann and Sherman, 1998; Schmahmann, 2000, 2004).

This theory of a Universal Cerebellar Transform may be more generalizable, however, and applicable to the striatum (caudate nucleus and putamen). Like the cerebellum, the complex cytoarchitecture of the caudate nucleus and putamen, including the striosome and matrix compartments, is essentially consistent throughout the striatum, although there is further neurochemical differentiation in the ventral parts of the caudate nucleus, putamen, and ventral stria-tum–nucleus accumbens (Holt et al., 1997).

With regard to thalamus, there may be an overarching commonality to thalamic function. Nadeau and Crosson (1997) suggest that thalamus is active in selecting specific neuronal networks in the cortical projection fields of thalamic nuclei, acting as a gate regulating attentional processes. By enabling specific types of working memory, for example, thalamus facilitates selective engagement in different cognitive processes. But the histology of thalamic nuclei, like the histology of the cerebral cortex, varies from one nucleus to another. In the same way, then, that the specific contribution, i.e., the proposed universal transform, of prefrontal cortical area 45 is likely to differ from that of area 39 in the parietal lobe, or from primary visual area 17, subdivisions of the medial dorsal thalamic nucleus are likely to contribute in a different manner than the pulvinar medialis nucleus, or the magnocellular layer of the lateral geniculate nucleus. Within an anatomically homogenous, but connectionally complex area, such as pulvinar medialis, for example, the transform should be constant throughout the nucleus and applied to anatomically segregated domains of cognitive processing. Such a nucleus-specific functional specialization has been proposed in the case of the pulvinar. Robinson (1993) hypothesizes that the pulvinar functions as an early center for the generation of visual salience by the suppression of noise and the enhancement of significant signals, enabling cortical shifts of attention and response specification. This notion, likely applicable to the pulvinar inferior that has strong connections with visual cortices (Yeterian and Pandya, 1997; Beck and Kaas, 1998), is consistent with our universal transform concept derived from a consideration of neuronal architecture and connections. The precise nature of the transforms of these thalamic nuclei, both at the level of the physiological transform, and the manner in which behavior is influenced, remains a matter of speculation.

12.3. Connections define behavior

The connections that link the multiple geographically distributed and architectonically defined nodes in the cerebral cortex and subcortical areas. Indeed, a cortical area can be defined by its pattern of subcortical and cortical connections. Whereas architecture may determine the transform of each node, connectional patterns are likely to define behavior. Each node in the distributed neural system are highly organized required to contribute its unique transform in order to support the ultimate behavior pattern, falling within any of the domains of sensorimotor processing, cognition or emotion. Further, each node may be engaged in a number of different domains of behavior, which are subserved by anatomically distinct subpopulations of neurons within the node. Clinical manifestations of lesions in the different nodes of the distributed neural system are therefore determined by two central factors – which node is interfered with, and which subpopulation of neurons within that node, or its connecting axons, are destroyed.

Centuries of investigation have led to an understanding of the functions subserved by different regions of the cerebral cortex, and cerebral cortical areas have come to be regarded as the parent nodes of the distributed networks. Based on the principles of organization of the cortical connections that we have outlined (Schmahmann and Pandya, 2006), a lesion of cortex, even if restricted exquisitely to the cortex, disconnects the cortex from all the other interconnected cortical and subcortical areas, because the neurons that provide the interconnecting axons are destroyed. The notion of Wernicke and Geschwind that a lesion in a cortical association area produces a disconnection syndrome by preventing cortical areas from communicating with each other is thus extended from the cortex to the subcortical areas. A cortical lesion may thus be expected to have a sizeable behavioral impact not only because its unique transform has been damaged, but also because widespread connections are affected throughout the particular distributed neural network for which that cortical area serves as the parent node.

The detailed clinical neurology of white matter lesions is beyond the scope of the present consideration, but lesions of white matter fiber pathways themselves produce clinical consequences. This point is illustrated by dementia that occurs in the setting of disseminated white matter disease (Filley, 2001), post-stroke language recovery that depends on involvement of the subcallosal fasciculus of Muratoff (Naeser et al., 1989), and parietal pseudothalamic pain from white matter lesions that disconnect SII cortex from thalamus (Schmahmann and Leifer, 1992). The clinical deficits from loss of the white matter tracts linking different nodes may differ from those following lesions of the cortex or sub-cortical nodes for a number of reasons. White matter lesions may disrupt information destined for more than one node; they may involve association, projection and striatal fibers; and they may affect more than one functional domain. Involvement of afferent versus efferent fibers may have clinical significance – striatal fibers are unidirectional from cortex to caudate–putamen; the middle cerebellar peduncle is essentially exclusively afferent from pons to cerebellum, whereas the superior cerebellar peduncle is predominantly efferent from cerebellum to the cerebral hemispheres; and the thalamic peduncles are bi-directional. Fiber tract disruptions are often incomplete by virtue of the anatomical arrangement of the pathways and the pathologic conditions that affect white matter, and the effects of partial versus complete disconnection are likely to be pertinent.

12.4. Topography of function in subcortical nodes – functional and clinical implications

The foregoing discussion sets the stage for considering the two principles that we postulate govern the clinical manifestations resulting from lesions of the subcortical nodes – the topography of the lesion within the node (i.e., the behavioral domain affected), and the transform of the affected node. This can be illustrated using selected clinical examples.

Motor impairment may result from focal lesions in the precentral gyrus, basal ganglia, thalamus and cerebellum. Further, the nature of the clinical impairment differs depending on the node affected. The clumsy or paralyzed arm results from stroke involving the upper extremity representation in the precentral gyrus (Takahashi et al., 2002; Castaldo et al., 2003; Maeder-Ingvar et al., 2005); dystonia and slowness of movement results from lesions of the motor regions of the putamen (Bhatia and Marsden, 1994); the motor cerebellum in the anterior lobe produces incoordination of movement (Holmes, 1917; Schmahmann, 2007); and the ventral lateral thalamic nucleus lesion may produce ataxic hemiparesis (Dejerine and Roussy, 1906; Caplan et al., 1988; Murthy, 1988). The most prolonged and unremitting paralysis occurs following hemispheric lesions (Shelton and Reding, 2001; Ng et al., 2007) that destroy not just corticospinal fibers, but the fiber tracts linking motor cortices with movement related neuronal populations in the thalamus and striatum as well (Schmahmann and Pandya, 2006). The clinical impairments following involvement of the different cortical and subcortical components of the motor system make the point that each node contributes in a unique manner to the motor program, and all contributing nodes are required to produce the normal behavior.

Cognitive impairments in the executive domain result from lesions of the dorsolateral prefrontal cortex (Stuss and Benson, 1984) and its interconnected areas in the dorsolateral caudate nucleus (Caplan et al., 1990), medial dorsal thalamic nucleus (Bogousslavsky et al., 1988), and crus I and II of cerebellum (Schmahmann and Sherman, 1998). Visual spatial impairments follow lesions of the posterior parietal cortex (Mesulam, 1981), the head of the caudate nucleus (Caplan et al., 1990), thalamus including the lateral posterior thalamic nucleus and pulvinar (Watson and Heilman, 1979), and the posterior lobe of the cerebellum (Schmahmann and Sherman, 1998). Disorders of affect, drive and motivation occur following lesions of the cingulate gyrus (Devinsky and Luciano, 1997), the ventral striatum (Levy and Dubois, 2006), the anterior thalamus (Graff-Radford et al., 1985; Lisovoski et al., 1993), and the cerebellar vermis (Pollack, 1997; Schmahmann and Sherman, 1998; Levisohn et al., 2000; Riva and Giorgi, 2000; Schmahmann et al., 2007c). Cognitive impairments following these subcortical lesions have been conceptualized as follows. The slowness of movement (bradykinesia), thought (bradyphrenia), and mnemonic retrieval that characterize lesions of the basal ganglia suggest that the striatal deficits impair the initiation of behavior, and the ability to chunk information into manageable quanta (Graybiel, 1998). The thalamus is thought to contribute a specific alerting or engagement response to the different domains of function to which it contributes. And the cerebellar role in automatization and optimizing behavior around a homeostatic baseline has been encapsulated in the dysmetria of thought theory – applying the same transform to cognition and emotion as it does to motor control (Schmahmann, 1991, 1996, 2000, 2004).

Thus, a cortical area can be defined by its pattern of subcortical (as well as cortical) connections; lesions of subcortical structures mimic deficits resulting from lesions of the cerebral cortex; and there are qualitative differences between the clinical manifestations of lesions in functionally related areas of different cortical and subcortical nodes.

12.5. The problem of segregated loops – the role of the cerebral cortex

Destruction of selected neuronal populations within subcortical nodes produces disorders of sensorimotor or higher order behavior because these lesions disrupt the interaction of domain-specific subcortical neuronal ensembles with the sensorimotor, association, or paralimbic cortical areas that govern these behaviors. But the existence of multiple parallel loops linking the cerebral cortex with subcortical nodes leads to a conceptual problem. Cortical connections with the striatum, thalamus and pontocerebellar system appear to be arranged in parallel, respecting domain specificity. There is no cross-modal communication within subcortical nodes (e.g., thalamic nuclei do not communicate with each other). Further, present evidence indicates that communication between subcortical nodes is also domain-specific, although we are mindful of possible exceptions as noted in the striatonigostriatal system (Haber et al., 2000).

Hence the difficulty. There is essentially no intranodal communication, and internodal communication is domain-specific; but complex behavior requires interactions between different functional domains. At some point in the anatomy of cognition, mood must inform movement, strategy must rely on memory, and so on. As emphasized early on by Franz Joseph Gall (1758–1828), Theodore Meynert (1833–1892), Wernicke and others, this is the role of the cerebral cortical association areas that “serve as the repository of past impressions arising through the afferent channels via the sensory spheres, and as the essential substratum of memory in integration or building of complex processes” (Paul Flechsig [1847–1929], translated and quoted in Barker, 1899, p. 1070). The problem of segregated loops is resolved by the cerebral cortex that, alone among the many nodes of the distributed neural system, has the privilege of facilitating the interaction and integration of information between multiple domains in a feedforward and feedback manner. This capability is facilitated by the association fiber pathways, which are themselves exclusive to the cerebral cortex.

12.6. Human brain connectivity

Knowledge of the anatomy of cortico-cortical and cortico-subcortical connections and fiber pathways of the brain is based largely on tract tracing studies in monkey. Clinical syndromes in patients are in good agreement with conclusions derived from these anatomical investigations, providing some confidence that the anatomical arrangements are likely to be generally preserved in the human. Thus, by their consistency with monkey anatomy, the clinical syndromes help clarify human brain hodology. The greater size, anatomic complexity, and functional elaboration of the human brain make it clear that this inter-species comparison may be erroneous, however, and there is a compelling need to study the anatomy and functional organization of the human brain directly, a challenge that has faced neuroscience since the ancient Greeks.

The ability of MRI to identify small, acute focal lesions, together with sophisticated tools of clinical neurology and neuropsychology, enables the structure–function correlations discussed in this analysis. The advent of diffusion tensor imaging (DTI, Basser et al., 1994) makes it possible to visualize fiber tracts in the human brain (Wedeen et al., 1995; Conturo et al., 1999; Catani et al., 2002) in a manner that may translate the findings from monkey to human. This technology is now being utilized to demonstrate some of the evolutionary principles described above, as in the case of the expansion in human of the prefrontal component of the cerebral peduncle fibers that contribute to the corticopontine pathways (Ramnani et al., 2006). Connectional anatomical investigations using DTI probabilistic mapping have also been initiated for the study of cortical connections with subcortical structures. Studies of thalamic interactions with the cerebral cortex report that a large medial, dorsal region of the thalamus had highly probable prefrontal and temporal connections; a ventral posterior thalamic region had strong probability of somatosensory connections; a lateral region had high probability of motor cortical connectivity; and a posterior region was connected to the posterior parietal cortex and extrastriate cortices (Behrens et al., 2003; Johansen-Berg et al., 2005). These findings are in general agreement with those of tracer studies in monkey, as discussed in detail above.

DTI tractography has also been performed in the analysis of human corticostriatal projections. This approach revealed that the dorsolateral prefrontal cortex is linked with the dorsal–posterior caudate nucleus, and the ventrolateral prefrontal cortex is linked with the ventral–anterior caudate nucleus. For the putamen, connections were defined between the supplementary motor area (SMA) and the dorsal–posterior putamen; the premotor area and the medial putamen; and the primary motor cortex and the lateral putamen (Leh et al., 2007). These findings complement those of an earlier investigation (Lehéricy et al., 2004) showing that the motor cortex and supplementary motor area (SMA) have connections with similar parts of the sensorimotor compartment of the human striatum, whereas the pre-SMA is linked with more rostral parts of the striatum, including the associative compartment. As in the case of the probabilistic DTI investigations of human thalamic connections, these striatal findings are also generally consistent with those derived from tract tracer injection studies in monkey.

The DTI demonstration of the fiber pathways and the connections they convey is presently far short of the details elucidated in the experimental animal. Diffusion tensor imaging cannot adequately demonstrate the anatomic reality of crossing fibers (Wedeen et al., 1995), a limitation that has been addressed with high angular resolution methods including diffusion spectrum imaging (Wedeen et al., 2005, 2008; Schmahmann et al., 2007b; Fig. 13), Q-ball (Tuch et al., 2002) and “HARDI”. The ultimate goal is to provide greater sophistication to anatomic analyses, in order to more closely approximate the details of fiber anatomy and connectional patterns achieved by neuronal tracers. The detection of white matter bundles by DSI has now been validated in the monkey (Schmahmann et al., 2007b). This approach may enable connectional neuroanatomical studies in the human brain that more closely approximate the detailed understanding derived from tract tracing studies in the experimental animal, a desirable goal for the field of neuroimaging in the coming years.

Fig. 13.

Figures illustrating the results of diffusion spectrum imaging (DSI) tractography of the centrum semiovale of monkey at intermediate (A) and high magnification (B). Commissural fibers (Comm) coursing towards the corpus callosum (orange) intersect with vertically oriented fibers of the subcortical bundle (SB) in the corona radiata (green), and with rostrocaudally oriented long association fibers (Assn, blue). Interpenetration of these fiber fascicles is evident (from Wedeen et al., 2008).

A novel approach to mapping connections in the human brain is resting state functional connectivity using MRI (fcMRI). This method detects temporal correlations in the spontaneous oscillations of the blood oxygenation level dependent (BOLD) signal between brain regions while subjects rest quietly in the scanner (Raichle et al., 2001; Greicius et al., 2003, 2008). By showing that specific groups of brain areas show greater activity during rest than during cognitive states, fcMRI has provided support for the existence of a baseline default mode of brain function that is suspended during specific goal-directed behaviors. It has also enabled studies of functional connectivity in distributed networks in the human brain (van de Ven et al., 2004; Margulies et al., 2007), including studies specifically of attentional networks in the frontoparietal and cingulo-opercular regions (Fair et al., 2007). Resting state fcMRI has been combined with DTI tractography to provide support for the notion that resting state functional connectivity indeed reflects structural connectivity (Greicius et al., 2008). Voxelwise regression analysis in an fcMRI study of the basal ganglia (Di Martino et al., 2008) provided support for the existence of striatal motor, cognitive and affective divisions, and for the presence of parallel loops as determined previously by anatomical studies in animals and functional imaging studies in humans. Further, fcMRI studies in patients with depression have identified heightened connectivity between the subgenual cingulate gyrus, implicated in depression (Mayberg et al., 2005), and the thalamus (Greicius et al., 2007). These findings from structural and functional MRI provide independent validation of the network model of brain function, and emphasize the degree to which subcortical structures are integral to this notion.

Finally, the use of clinical case material, despite all its limitations, remains an invaluable tool for the detailed evaluation of how structure relates to function. Discriminative clinical correlation with cortical, subcortical, and white matter lesions that are defined precisely by structural and functional MRI has provided valuable insights into the organization and function of the human nervous system. It has also reaffirmed notions of the distributed neural circuits and disconnection syndromes, and deepened the interdependence of clinical neuroscience and connectional neuroanatomy.

13. Conclusion

We have reviewed anatomical, clinical, and imaging data implicating the thalamus, basal ganglia and cerebellum in cognition and emotion, in addition to their roles in motor control. We consider these clinical phenomena as disconnection syndromes, extending the earlier notions of Wernicke and Geschwind to the subcortical nodes. We introduce a new theoretical framework in which to consider the distributed neural systems, based on general, and specific, principles of anatomical organization, and on the cytoarchitecture of the nodes that comprise these systems. This model enables the formulation and testing of future hypotheses in imaging and clinical studies.

List of abbreviations+

AD

anterior dorsal thalamic nucleus

AM

anteromedial thalamic nucleus

AS

arcuate sulcus

Assn

association fibers

AV

anteroventral thalamic nucleus

CC

corpus callosum

CCAS

cerebellar cognitive affective syndrome

Cd

caudate nucleus

CeM

central medial thalamic nucleus

CL

central lateral thalamic nucleus

Cl

claustrum

Cld

central lateral dorsal thalamic nucleus

CM

centromedian thalamic nucleus

Comm

commissural fibers

CS

central sulcus

Csl

centralis superior lateralis thalamic nucleus

DM

medial dorsal thalamic nucleus

DSI

diffusion spectrum magnetic resonance imaging

DTI

diffusion tensor magnetic resonance imaging

DWI

diffusion weighted imaging

EC

external capsule

F

fornix

GM

medial geniculate nucleus

GMpc

medial geniculate nucleus, parvicellular part

GPe

globus pallidus, external segment

GPi

globus pallidus, internal segment

H

habenula

IC

internal capsule

ICa

internal capsule, anterior limb

ICp

internal capsule, posterior limb

IL

intralaminar thalamic nucleus

IPS

inferior parietal sulcus

IQ

intelligence quotient

L/Li

limitans thalamic nucleus

LD

lateral dorsal thalamic nucleus

LF

lateral fissure

LGN/LGd/LGB

lateral geniculate thalamic nucleus

LP

lateral posterior thalamic nucleus

LS

lunate sulcus

M1

primary motor cortex

MB

Muratoff bundle (subcallosal fasciculus of)

MD

medial dorsal thalamic nucleus

MDdc

medial dorsal thalamic nucleus, densocellular part

MDmc

medial dorsal thalamic nucleus, magnocellular part

MDmf

medial dorsal thalamic nucleus, multiform part

MDpc

medial dorsal thalamic nucleus, parvicellular part

MGN/MG

medial geniculate thalamic nucleus

MRI

magnetic resonance imaging

MRI

magnetic resonance imaging

MTT

mamillothalamic tract

P.Comm

posterior communicating artery

P1,P2

first, second divisions of the posterior cerebral artery

Pcn

paracentral thalamic nucleus

Pf

parafascicular thalamic nucleus

PI

inferior pulvinar thalamic nucleus

PICA

posterior inferior cerebellar artery

PL/Pll

lateral pulvinar thalamic nucleus

PM/Plm

medial pulvinar thalamic nucleus

PMV

ventral premotor area

PO/Pla

anterior pulvinar (pulvinar oralis) thalamic nucleus

PS

principal sulcus

Pt

paratenial thalamic nucleus

Pv

paraventricular thalamic nucleus

R

reticular thalamic nucleus

Re

reunions thalamic nucleus

RN

red nucleus

SB

subcortical bundle

SG, Sg

suprageniculate thalamic nucleus

SII

second somatosensory area

SLF

superior longitudinal fasciculus

Sm

stria medullaris

SN

substantia nigra

ST

subthalamic nucleus

StB

striatal bundle

STS

superior temporal sulcus

Th

thalamus

UF

uncinate fasciculus

VA

ventral anterior thalamic nucleus

VL

ventral lateral thalamic nucleus

VLa

ventral lateral anterior nucleus (Jones) = VLo (Olszewski)

VLc

ventral lateral thalamic nucleus, pars caudalis (Olszewski) = VLp (dorsal part) [Jones]

VLo

ventral lateral thalamic nucleus, pars oralis (Olszewski) = VLa (Jones)

VLp

ventral lateral posterior nucleus (Jones)

VLps

ventral lateral thalamic nucleus, pars postrema (Olszewski) = VLp (posterodorsal part) [Jones)]

VM

ventromedial thalamic nucleus

VP

ventral posterior thalamic nucleus

VPI

ventral posterior inferior thalamic nucleus

VPL

ventral posterolateral thalamic nucleus

VPLa

ventral posterior lateral nucleus, anterior division (Jones) = VPLc (Olszewski)

VPLc

ventral posterolateral thalamic nucleus, pars caudalis (Olszewski) = VPLa (Jones)

VPLo

ventral posterolateral thalamic nucleus, pars oralis (Olszewski) = VLp (ventral part) [Jones]

VPM

ventral posteromedial thalamic nucleus

VPMpc

ventral posteromedial thalamic nucleus, parvicellular part

ZI

zona incerta

⁺: Comparisons of nomenclature for ventral tier thalamic nuclear according to Olszewski (1952) and Jones (1997)