Brain regions: cerebral cortex functional areas, basal ganglia, cerebellum, and limbic system

Intuition Beginner

The brain has specialized regions, each responsible for different jobs. The cerebral cortex is the wrinkled outer layer that handles thinking, voluntary movement, and sensation. Deep inside the brain, three other structures work behind the scenes. The basal ganglia help select and coordinate movements. The cerebellum fine-tunes balance, timing, and coordination. The limbic system processes emotions and forms memories.

These regions do not work in isolation. They are like sections of an orchestra: each has its own part, but they must play together to produce coherent behaviour. Damage to one region disrupts the whole performance in predictable ways — Parkinson's disease (basal ganglia) causes tremor and rigidity, cerebellar damage causes clumsy uncoordinated movements, and limbic damage affects memory and emotional regulation.

Visual Beginner

The cerebral cortex is divided into four pairs of lobes. The frontal lobe (front) controls voluntary movement, decision-making, and planning. The parietal lobe (top-rear) processes touch and body position. The temporal lobe (sides) handles hearing and language comprehension. The occipital lobe (back) processes vision.

Worked example Beginner

A patient has a stroke that damages a small area in the left frontal lobe. He understands everything said to him but cannot produce fluent speech. His words come out as effortful, telegraphic fragments ("dog...walk...park"). This is Broca's aphasia, caused by damage to Broca's area in the left inferior frontal gyrus. The comprehension network (Wernicke's area, in the left temporal lobe) is intact, but the speech production engine is broken.

Contrast this with damage to Wernicke's area in the left temporal lobe: the patient speaks fluently with normal rhythm and grammar, but the words are meaningless jumbles ("I went to the fliggle and rambled the spools"). Comprehension is lost. These two patterns illustrate the functional specialisation of cortical regions: different areas handle different aspects of language.

Check your understanding Beginner

Formal definition Intermediate+

Cerebral cortex: lobes and functional areas

The cerebral cortex is a ~2-4 mm thick sheet of grey matter divided into four lobes per hemisphere. Function is localised to specific cortical regions:

Frontal lobe:

• Primary motor cortex (M1) — precentral gyrus, Brodmann area 4. Somatotopically organised with a distorted body map (motor homunculus): disproportionate representation of the hands, face, and tongue. Controls voluntary movement of contralateral body parts.
• Premotor cortex — area 6. Programs movement sequences before execution.
• Supplementary motor area (SMA) — medial area 6. Involved in planning bilateral and sequential movements.
• Broca's area — areas 44/45 in the left inferior frontal gyrus. Speech production. Damage causes Broca's aphasia (non-fluent, effortful speech with preserved comprehension).
• Prefrontal cortex — anterior frontal lobe. Executive function: working memory, planning, decision-making, impulse control, social behaviour. Last cortical region to myelinate (continuing into the mid-twenties).

Parietal lobe:

• Primary somatosensory cortex (S1) — postcentral gyrus, areas 3, 1, 2. Receives input from VPL/VPM of the thalamus. Somatotopically organised (sensory homunculus) mirroring the motor homunculus. Processes touch, pressure, vibration, temperature, and proprioception from the contralateral body.
• Posterior parietal cortex — areas 5, 7. Integrates somatosensory, visual, and motor information for spatial awareness, attention, and sensorimotor transformation (converting sensory coordinates into motor plans).

Temporal lobe:

• Primary auditory cortex (A1) — areas 41/42, Heschl's gyrus. Tonotopically organised: different frequencies map to different cortical positions.
• Wernicke's area — area 22, posterior superior temporal gyrus (left hemisphere). Language comprehension. Damage causes Wernicke's aphasia (fluent, meaningless speech with impaired comprehension).
• Medial temporal lobe — hippocampus and surrounding cortex. Memory formation (declarative memory).

Occipital lobe:

• Primary visual cortex (V1) — area 17, calcarine sulcus. Retinotopically organised. Receives input from the lateral geniculate nucleus (LGN) of the thalamus.
• Visual association cortices (V2-V5) — areas 18, 19 and beyond. Process colour, motion, form, and depth. Dorsal stream ("where/how") projects to parietal cortex; ventral stream ("what") projects to temporal cortex.

Association areas occupy the remaining cortex and integrate information across modalities. Multimodal association cortex in the prefrontal, posterior parietal, and temporal lobes underpins higher cognitive functions.

Basal ganglia

The basal ganglia are a group of deep nuclei involved in movement selection, initiation, habit learning, and reward-based decision-making. Their major components:

• Caudate nucleus and putamen — together form the dorsal striatum, the input nucleus. Receive excitatory (glutamatergic) input from the cerebral cortex.
• Globus pallidus — internal segment (GPi) and external segment (GPe). GPi is a major output nucleus.
• Subthalamic nucleus (STN) — receives input from GPe and cortex, projects to GPi.
• Substantia nigra — pars compacta (SNc) provides dopaminergic input to the striatum; pars reticulata (SNr) is a GPi-homologous output nucleus.

The basal ganglia operate through two opposing pathways originating in the striatum:

Direct pathway (facilitates movement): Cortex $\to$ striatum (D1-expressing medium spiny neurons) $\to$ GPi/SNr $\to$ thalamus $\to$ cortex. Dopamine from SNc excites D1 neurons, facilitating this pathway. The striatum inhibits GPi, which normally inhibits the thalamus — so striatal activity disinhibits the thalamus, promoting movement.

Indirect pathway (suppresses movement): Cortex $\to$ striatum (D2-expressing medium spiny neurons) $\to$ GPe $\to$ STN $\to$ GPi/SNr $\to$ thalamus $\to$ cortex. Dopamine from SNc inhibits D2 neurons, suppressing this pathway. The net effect is increased GPi inhibition of the thalamus, suppressing unwanted movements.

The balance between these pathways allows the basal ganglia to act as a gate that selects appropriate motor programs while suppressing competing ones.

Cerebellum

The cerebellum ("little brain") sits below the occipital lobe, posterior to the brainstem. It contains roughly 69 billion granule cells — more neurons than the rest of the brain combined. Anatomically divided into:

• Anterior lobe — primarily involved in posture and gait.
• Posterior lobe — involved in fine motor coordination and motor learning.
• Flocculonodular lobe — vestibulocerebellum, involved in balance and eye movements.

The cerebellar cortex has three layers with a stereotyped circuit:

1. Granule cells (granular layer) — receive mossy fibre input from pontine nuclei (carrying copies of cortical motor commands) and spinal cord (proprioceptive feedback). Their axons ascend as parallel fibres in the molecular layer.
2. Purkinje cells (Purkinje layer) — receive convergent input from $\sim$200,000 parallel fibres (weak excitatory synapses) and a single climbing fibre from the inferior olive (powerful excitatory synapse). Purkinje cells are the sole output of cerebellar cortex and are inhibitory (GABAergic).
3. Purkinje cells project to the deep cerebellar nuclei (fastigial, interpositus, dentate), which then send excitatory output to the thalamus (dentate $\to$ VL thalamus $\to$ motor cortex), the red nucleus, and vestibular nuclei.

The cerebellum functions as an error-correction device: it compares the intended movement (cortical motor command copy via mossy fibres) with the actual movement (sensory feedback), computes the error, and sends corrective signals back to motor cortex. The climbing fibre is thought to carry the error signal that drives long-term depression (LTD) at parallel fibre-Purkinje synapses, implementing supervised motor learning (Marr-Albus-Ito model).

Limbic system

The limbic system is a set of structures on the medial surface of the brain involved in emotion, motivation, memory, and olfaction. Core components:

• Hippocampus — medial temporal lobe. Critical for formation of new declarative memories (episodic and semantic). Consolidates short-term memories into long-term storage in the neocortex. Damage produces anterograde amnesia (inability to form new memories), as in patient H.M.
• Amygdala — anterior medial temporal lobe, at the tip of the hippocampus. Processes fear, threat detection, and emotional learning. Assigns emotional valence to sensory stimuli. Central to fear conditioning (aversive learning).
• Cingulate cortex — on the medial wall of the hemisphere, arching above the corpus callosum. Anterior cingulate cortex (ACC) is involved in error detection, conflict monitoring, pain affect, and motivation. Posterior cingulate cortex is a hub of the default mode network.
• Parahippocampal gyrus — surrounds the hippocampus. Involved in spatial memory and scene recognition.
• Fornix — white matter tract connecting the hippocampus to the mammillary bodies and septal nuclei.
• Mammillary bodies — hypothalamus. Involved in memory circuits (Papez circuit).

Key mechanism Intermediate+

Basal ganglia direct and indirect pathways: dopamine balance

The direct and indirect pathways of the basal ganglia operate as a push-pull system controlling movement selection. The mechanism hinges on the dual effect of dopamine on two populations of striatal neurons:

Direct pathway (movement facilitation):

1. Cortical glutamatergic input excites striatal neurons expressing D1 dopamine receptors.
2. SNc dopamine excites these D1 neurons via $G_s$-coupled signalling (increases cAMP, enhances excitability).
3. D1 striatal neurons project directly to GPi/SNr, releasing GABA and inhibiting the GPi/SNr output.
4. GPi/SNr normally tonically inhibit the thalamus (ventral anterior/ventral lateral nuclei) via GABA release.
5. Inhibiting GPi/SNr disinhibits the thalamus, allowing thalamocortical excitation of motor cortex.
6. Net effect: the selected motor program is facilitated.

Indirect pathway (movement suppression):

1. Cortical glutamatergic input excites striatal neurons expressing D2 dopamine receptors.
2. SNc dopamine inhibits these D2 neurons via $G_i$-coupled signalling (decreases cAMP, reduces excitability).
3. D2 striatal neurons inhibit GPe (external globus pallidus).
4. GPe normally inhibits the subthalamic nucleus (STN).
5. Inhibiting GPe disinhibits the STN, allowing STN to excite GPi/SNr (glutamatergic projection).
6. Increased GPi/SNr output increases thalamic inhibition, suppressing motor cortex.
7. Net effect: competing motor programs are suppressed.

The dual-receptor arrangement means that dopamine from SNc simultaneously facilitates the direct pathway (exciting D1) and inhibits the indirect pathway (inhibiting D2). Both effects converge on the same outcome: promoting the selected movement while suppressing alternatives.

Clinical consequences:

• Parkinson's disease — degeneration of dopaminergic neurons in SNc reduces dopamine input to striatum. Both effects are lost: the direct pathway is weakened (less movement facilitation) and the indirect pathway is overactive (excessive movement suppression). Result: bradykinesia (slowness), rigidity, resting tremor, and postural instability.
• Huntington's disease — degeneration of striatal neurons, preferentially the D2-expressing population of the indirect pathway. The indirect pathway is weakened, reducing movement suppression. Result: chorea (involuntary jerky movements), along with cognitive decline and psychiatric symptoms.

Exercises Intermediate+

Cerebellar circuitry and motor learning Master

Cerebellar microcircuit in detail

The cerebellar cortex has one of the most regular and well-characterised circuits in the brain. Its three-layered structure implements a massive expansion of input dimensionality followed by convergent readout:

Input pathways:

• Mossy fibres originate from pontine nuclei (carrying cortical motor command copies), spinal cord (proprioceptive feedback), and vestibular nuclei. Each mossy fibre terminates in a glomerulus in the granular layer, where it excites an average of 4-5 granule cells. There are approximately $10^5$ mossy fibres and $6.9\times 10^{10}$ granule cells, producing an expansion factor of roughly $500$ (the "sparse coding" strategy: each granule cell is active only briefly and rarely, at firing rates of $\sim 0.5$ Hz, but the population as a whole can represent high-dimensional input patterns).

• Climbing fibres originate exclusively from the inferior olive. Each Purkinje cell receives input from exactly one climbing fibre (one of the most specific connections in the CNS). Climbing fibre activation produces a powerful, all-or-none complex spike in the Purkinje cell: a large calcium-dependent depolarisation followed by a burst of 2-5 spikelets, readily distinguishable from the simple spikes produced by parallel fibre input.

Cerebellar cortex circuit:

• Granule cell axons ascend into the molecular layer and bifurcate into parallel fibres running mediolaterally for several millimetres. Each parallel fibre passes through the dendritic trees of hundreds to thousands of Purkinje cells, making one or two synapses on each. Each Purkinje cell receives input from $\sim 200,000$ parallel fibres but only one climbing fibre.
• Interneurons of the molecular layer (stellate cells, basket cells) provide feedforward inhibition onto Purkinje cells, sharpening the spatial and temporal profile of parallel fibre activation. Basket cells produce powerful perisomatic inhibition that can silence Purkinje cell output entirely. Golgi cells in the granular layer provide feedback inhibition onto granule cells, controlling the gain of the mossy fibre $\to$ granule cell transformation.

Output: Purkinje cells are GABAergic and project inhibitorily to the deep cerebellar nuclei (and to the vestibular nuclei for the flocculonodular lobe). The deep nuclei also receive direct excitatory collaterals from mossy and climbing fibres, providing a tonic excitatory drive that is sculpted by Purkinje cell inhibition.

Marr-Albus-Ito model of motor learning

The dominant theory of cerebellar motor learning, developed across four decades (Marr 1969, Albus 1971, Ito 1982, and subsequent experimental confirmation):

1. Mossy fibres carry the context signal (the current state of the body and the intended movement). The granule cell layer expands this into a sparse, high-dimensional representation.
2. Climbing fibres carry the error signal (the discrepancy between intended and actual movement, computed in the inferior olive from comparisons of motor commands and sensory feedback).
3. Learning rule: When a climbing fibre fires (error detected), it triggers long-term depression (LTD) at the parallel fibre synapses that were active immediately before the error. LTD reduces the strength of those synapses, decreasing the Purkinje cell's response to the context that produced the error.
4. Result: The next time the same movement context arises, the weakened parallel fibre synapses produce less Purkinje cell inhibition of the deep nuclei, allowing more excitatory output to the thalamus and motor cortex. The movement is corrected.

This is a supervised learning algorithm: the climbing fibre is the teacher, the parallel fibres are the student inputs, and the Purkinje cell output is the prediction. The algorithm converges because errors drive synaptic changes that reduce future errors.

Experimental support: Ito and colleagues demonstrated LTD at parallel fibre-Purkinje synapses in vitro (1982). Gilbert and Thach (1977) showed that motor learning in awake monkeys is accompanied by decreased Purkinje cell simple-spike firing. Optogenetic activation of climbing fibres in behaving animals (medial vestibular nucleus and VOR adaptation studies) can substitute for natural error signals and drive learned changes in motor output.

Cerebellar cognitive functions

Beyond motor control, the cerebellum contributes to cognitive and affective processing. The posterior lobule regions (Crus I, Crus II, lobules VI-VII) are reciprocally connected with prefrontal and parietal association cortices via the dentate nucleus and thalamus. Functional imaging and lesion studies implicate the cerebellum in:

• Language processing — verbal fluency, grammar, and the timing of speech.
• Working memory — particularly the rehearsal and manipulation of information.
• Sequence learning — the implicit learning of serial patterns (e.g., serial reaction time tasks).
• Social cognition — theory of mind and the prediction of others' behaviour.
• Emotional regulation — connections to limbic structures via the fastigial nucleus.

The "dysmetria of thought" hypothesis (Schmahmann 1991) proposes that cerebellar damage produces a cognitive-affective syndrome analogous to motor ataxia: imprecise, poorly coordinated thinking rather than motor incoordination. The cerebellar cognitive affective syndrome (CCAS), described after cerebellar stroke and tumour resection, includes deficits in executive function, spatial cognition, personality change, and language dysfunction.

Hippocampal circuitry, place cells, and memory consolidation Master

Hippocampal anatomy

The hippocampus is a curved structure in the medial temporal lobe with a characteristic laminar organisation. Its principal circuit (the trisynaptic circuit) flows unidirectionally:

1. Entorhinal cortex (layer II) $\to$ dentate gyrus via the perforant path (excitatory, glutamatergic).
2. Dentate gyrus granule cells $\to$ CA3 pyramidal cells via mossy fibres (powerful, sparse, one-to-few connectivity).
3. CA3 $\to$ CA1 via Schaffer collaterals (excitatory, with extensive recurrent collaterals within CA3 forming an autoassociative network).
4. CA1 $\to$ subiculum $\to$ entorhinal cortex (layer V) — completing the loop and projecting back to the neocortex.

Each stage performs a distinct computation. The dentate gyrus performs pattern separation: it orthogonalises overlapping inputs, mapping similar experiences to distinct sparse representations (aided by adult neurogenesis of dentate granule cells). CA3 performs pattern completion: its recurrent collateral network can retrieve a complete memory from a partial cue. CA1 performs comparison between the retrieved CA3 pattern and the current entorhinal input, detecting novelty.

Place cells and spatial navigation

O'Keefe and Dostrovsky (1971) discovered that hippocampal CA1 pyramidal cells fire selectively when a rat occupies a specific location in its environment. Each place cell has a place field — a restricted region of space where its firing rate increases sharply. The population of place cells collectively encodes the animal's position, and the hippocampus is therefore described as a cognitive map (O'Keefe and Nadel 1978).

Properties of place cells:

• Remapping — when the environment changes (different room, different wall colour, different shape), place fields remap: cells that were active in the old environment may become silent, and previously silent cells may acquire new place fields. This reflects the pattern separation function of the hippocampal circuit.
• Phase precession — as a rat traverses a place field, the cell's spikes occur at progressively earlier phases of the theta rhythm (the 6-10 Hz oscillation that dominates hippocampal EEG during locomotion and exploration). This means that within a single theta cycle, cells representing locations ahead of the animal fire in sequence, providing a temporal code for trajectory.
• Goal-direction — place fields shift toward goal locations in navigation tasks, reflecting the influence of reward and behavioural relevance on spatial coding.

The discovery of grid cells in the medial entorhinal cortex (Hafting et al. 2005) provided the input to the hippocampal spatial system. Grid cells fire at multiple locations arranged in a regular hexagonal grid across the environment, providing a metric for spatial navigation. Grid cells, together with head-direction cells, border cells, and speed cells in the entorhinal cortex, supply the hippocampus with a rich representation of position, heading, and environmental geometry.

Memory consolidation: systems-level replay

The hippocampus does not store memories permanently. It acts as a fast-learning system that binds the distributed cortical representations of an experience into a coherent engram, then gradually transfers that engram to the neocortex through a process called systems consolidation.

The key mechanism is sharp-wave ripple oscillations (SWR, 100-250 Hz) that occur during slow-wave sleep and quiet wakefulness. During SWR events, large populations of hippocampal neurons fire in sequences that replay the spatial-temporal patterns of prior experience — but at compressed timescales (a 10-second experience is replayed in $\sim 100$ ms). This replay drives synaptic plasticity in the neocortex, strengthening the direct cortico-cortical connections that will eventually support memory retrieval without hippocampal involvement.

The standard consolidation model (Squire and Alvarez 1995) posits that recent memories are hippocampus-dependent, while remote memories (weeks to years old) become hippocampus-independent as consolidation transfers them to the neocortex. This explains H.M.'s temporally graded retrograde amnesia: memories formed shortly before his surgery were lost (not yet consolidated), while remote childhood memories were intact (already stored in neocortex).

Amygdala, fear conditioning, and emotional learning Master

Amygdala nuclei and circuitry

The amygdala is an almond-shaped cluster of nuclei in the anterior medial temporal lobe. Its major subdivisions:

• Basolateral amygdala (BLA) — receives sensory inputs from thalamus (rapid, coarse, low-road) and from sensory cortex (slower, detailed, high-road). Contains excitatory pyramidal-like neurons and inhibitory interneurons. The site of plasticity for fear learning.
• Central nucleus (CeA) — the output nucleus. Projects to hypothalamus (autonomic responses), periaqueductal grey (freeze behaviour, pain modulation), and brainstem nuclei (startle, hormone release). Organised into medial (CeM) and lateral (CeL) subdivisions.

Fear conditioning

Pavlovian fear conditioning is the most thoroughly characterised form of emotional learning. In the standard paradigm, a neutral stimulus (a tone, the conditioned stimulus or CS) is paired with an aversive stimulus (a footshock, the unconditioned stimulus or US). After one or a few pairings, the CS alone elicits a constellation of fear responses: freezing (cessation of movement), increased heart rate and blood pressure, startle potentiation, stress hormone release (corticotropin-releasing hormone, adrenocorticotropic hormone, cortisol), and analgesia.

The neural circuit:

1. The CS (tone) activates the auditory thalamus (medial geniculate nucleus), which projects directly to the BLA.
2. The US (shock) activates somatosensory thalamus and brainstem nociceptive pathways, which also project to the BLA.
3. In the BLA, convergent CS and US input drives Hebbian long-term potentiation (LTP) at CS-input synapses onto BLA pyramidal neurons. After LTP, the CS alone is sufficient to strongly activate these neurons.
4. BLA neurons project to CeA, which then activates the hypothalamus (autonomic fear responses), the periaqueductal grey (freezing behaviour), and the brainstem (startle reflex potentiation, hormone release).

Extinction is the reduction of conditioned fear by repeated CS presentation without the US. Extinction does not erase the original fear memory — it creates a new inhibitory memory that suppresses the fear response. Evidence: the fear response returns after a delay (spontaneous recovery), after a context change (renewal), or after an unpredictable stressor (reinstatement). Extinction depends on the infralimbic cortex (a subdivision of the medial prefrontal cortex), which projects to the BLA and intercalated cell clusters (GABAergic interneurons that inhibit CeA output).

Amygdala and anxiety disorders

Hyperactivity of the amygdala is a consistent finding in anxiety disorders, PTSD, and specific phobias. The amygdala's rapid, subcortical processing pathway allows threat responses to begin before conscious awareness — an adaptive mechanism for survival that becomes pathological when the threat detection system is miscalibrated. Treatment approaches that target amygdala circuits include:

• Exposure therapy — based on extinction principles. Repeated, controlled exposure to the feared stimulus (without the feared outcome) engages the infralimbic-amygdala inhibitory circuit.
• Cognitive behavioural therapy (CBT) — engages prefrontal cortex (particularly the ventromedial prefrontal cortex) to provide top-down regulation of amygdala reactivity.
• Pharmacological augmentation — D-cycloserine (a partial NMDA receptor agonist) enhances extinction learning when administered before exposure therapy, by facilitating LTP in the amygdala-prefrontal circuit.

Prefrontal cortex, executive function, and neuroplasticity Master

Prefrontal cortex organisation

The prefrontal cortex (PFC) comprises the anterior portion of the frontal lobes anterior to the premotor cortex. It is disproportionately large in humans ($\sim 30\%$ of cortical surface area, compared to $\sim 7\%$ in the cat and $\sim 3.5\%$ in the rat). It is the last cortical region to mature: synaptic pruning and myelination continue into the mid-twenties, which explains the protracted development of executive functions during adolescence and early adulthood.

The PFC is divided into three major regions:

Dorsolateral PFC (dlPFC, areas 9, 46):

• Working memory — maintaining and manipulating information online.
• Cognitive flexibility — switching between tasks, rules, or mental sets.
• Planning — organising multi-step sequences toward a goal.
• Abstract reasoning — forming and testing hypotheses.

Ventromedial PFC (vmPFC, areas 11, 12, 25, 32):

• Emotional decision-making — integrating emotional signals (from amygdala) into cost-benefit analyses.
• Social cognition — evaluating social norms, trust, and fairness.
• Extinction of conditioned fear — inhibitory control over amygdala output.
• Somatic marker hypothesis (Damasio) — gut feelings guide decisions via vmPFC connections to insula and amygdala.

Orbitofrontal cortex (OFC, areas 10-14):

• Reinforcement learning — updating the value of stimuli based on outcomes. Reversal learning (flexibly switching responses when reward contingencies change).
• Representing expected value and risk.
• Damage produces disinhibition, impulsivity, and poor social judgement.

Executive function and prefrontal control

Executive function is the set of cognitive processes that enable goal-directed behaviour. It depends on the PFC's ability to maintain representations of goals, rules, and context in working memory while inhibiting prepotent (automatic, habitual) responses. The key operations:

1. Inhibitory control — suppressing automatic responses. Measured by the Stroop task (reading colour words printed in incongruent colours), the Go/No-Go task, and the antisaccade task. Implemented by dlPFC networks that bias processing away from prepotent response pathways.

2. Working memory — actively maintaining and updating information. Baddeley's model comprises the phonological loop (verbal), the visuospatial sketchpad (spatial), the episodic buffer (integrative), and the central executive (attentional control). Persistent neural activity in dlPFC neurons (up states maintained by recurrent excitatory networks and NMDA receptor activation) is the neural substrate of working memory.

3. Cognitive flexibility — switching mental sets. Measured by the Wisconsin Card Sorting Test (sorting by one rule, then switching to another). Impaired after dlPFC damage and in schizophrenia.

4. Planning and problem-solving — the Tower of London/Hanoi tasks require multi-step planning. Requires dlPFC, particularly the left hemisphere.

Neuroplasticity and stroke recovery

The adult brain retains the capacity for structural and functional reorganisation (neuroplasticity) throughout life. The major mechanisms:

Synaptic plasticity: LTP and LTD modify the strength of existing synapses. This is the cellular basis of learning and memory, occurring at glutamatergic synapses throughout the cortex and hippocampus. NMDA receptor-dependent LTP requires coincident presynaptic glutamate release and postsynaptic depolarisation, allowing calcium influx through the NMDA channel. Calcium activates CaMKII, which phosphorylates AMPA receptors (increasing conductance) and triggers AMPA receptor insertion into the postsynaptic membrane.

Cortical map reorganisation: Sensory and motor cortical maps are dynamic. After amputation of a limb, the cortical territory previously devoted to that limb is invaded by representations of adjacent body parts (e.g., face representation expands into the hand area after arm amputation). This produces the phenomenon of referred sensation: touching the face is perceived as touch on the phantom limb. Similar reorganisation occurs with skill acquisition: string players have expanded cortical representation of the left hand (fingering hand) in somatosensory cortex.

Stroke recovery exploits neuroplasticity. After focal cortical damage (most commonly middle cerebral artery infarct affecting motor cortex and internal capsule), recovery proceeds through several stages:

1. Acute phase (days): Resolution of oedema and diaschisis (depression of activity in regions connected to the damaged area but not directly injured). Some immediate improvement.
2. Early subacute phase (weeks): Perilesional cortex (tissue surrounding the infarct) becomes hyperexcitable due to reduced GABAergic inhibition and upregulation of growth-promoting genes. Intensive rehabilitation (repetitive task practice, constraint-induced movement therapy) drives use-dependent plasticity, strengthening alternative pathways.
3. Late subacute and chronic phase (months): Structural changes in surviving pathways — axonal sprouting from the contralesional hemisphere, strengthening of ipsilesional corticoreticular and propriospinal pathways, and recruitment of supplementary motor areas. The extent of recovery depends on the size and location of the lesion, the intensity of rehabilitation, and the age of the patient.

Stroke localisation uses the functional anatomy of brain regions to infer lesion location from clinical deficits:

• Left MCA territory: right hemiparesis (motor cortex/internal capsule), right hemisensory loss (somatosensory cortex), Broca's aphasia (left inferior frontal), or Wernicke's aphasia (left superior temporal), depending on the specific branches affected.
• Right MCA territory: left hemiparesis, left hemisensory loss, left hemispatial neglect (right parietal — neglect is more common and severe after right parietal damage than left, because the right hemisphere attends to both visual fields while the left primarily attends to the right).
• PCA territory: visual field deficits (homonymous hemianopia from V1 damage), memory impairment (medial temporal lobe including hippocampus), or alexia without agraphia (left PCA affecting the splenium of the corpus callosum and left occipital lobe — the patient cannot read because visual input from the right hemisphere cannot reach the left angular gyrus for language processing, but can write because the left hemisphere language areas are intact).

Connectomics and large-scale brain networks Master

The connectome

The human connectome (Sporns, Tononi, and Kotter 2005) is the complete map of neural connections in the brain, described as a weighted, directed graph with nodes (brain regions, defined by a parcellation) and edges (white-matter connections, measured by diffusion tensor imaging or inferred from functional correlation). The Human Connectome Project (Van Essen et al. 2013) provided the first large-scale, high-resolution dataset of structural and functional connectivity in healthy adults.

Key graph-theoretic properties of the human connectome:

• Small-world architecture — high clustering coefficient (neighbours of a node tend to be connected to each other) and short characteristic path length (any two nodes are linked by a short chain of connections). This architecture balances local specialisation (segregation) with global integration.
• Hub nodes — a small number of regions have disproportionately many connections (high degree centrality and betweenness centrality). Hubs include the precuneus, posterior cingulate cortex, medial prefrontal cortex, thalamus, and insula. These are the integration centres of the brain.
• Rich-club organisation — hub nodes preferentially connect to other hub nodes, forming a dense "rich club" backbone. This backbone provides the high-bandwidth communication infrastructure that integrates specialised local processing.

Default mode network

The default mode network (DMN) is a set of brain regions that are consistently active at rest (during passive, internally-directed cognition) and deactivated during demanding external tasks. DMN nodes include:

• Medial prefrontal cortex (self-referential processing, mentalising).
• Posterior cingulate cortex / precuneus (episodic memory retrieval, self-projection).
• Lateral temporal cortex (semantic memory).
• Angular gyrus (attention, episodic memory).

The DMN was discovered by Raichle and colleagues (2001) through PET and fMRI meta-analysis showing that a specific set of regions consistently decreased their activity during goal-directed tasks compared to rest. The DMN is anti-correlated with the task-positive network (dorsal attention network, frontoparietal control network), which activates during externally-directed attention and cognitive control.

DMN dysfunction is implicated in several disorders: Alzheimer's disease (DMN nodes are the first regions to accumulate amyloid-beta and show metabolic decline), depression (hyperactive DMN, particularly the subgenual anterior cingulate, correlates with rumination), schizophrenia (abnormal DMN dynamics), and ADHD (reduced DMN deactivation during tasks, producing attentional lapses).

Parkinson's disease: pharmacotherapy and deep brain stimulation

The standard treatment for Parkinson's disease addresses the dopamine deficit through two approaches:

L-DOPA (levodopa): The immediate metabolic precursor of dopamine. Administered orally with a peripheral DOPA decarboxylase inhibitor (carbidopa or benserazide) to prevent peripheral conversion. L-DOPA crosses the blood-brain barrier via large neutral amino acid transporters and is converted to dopamine by central DOPA decarboxylase. It remains the most effective symptomatic treatment for Parkinson's disease. Long-term complications include:

• Motor fluctuations — "wearing off" (the dose effect shortens over years, producing predictable periods of immobility at end-of-dose).
• Dyskinesias — involuntary choreiform movements occurring at peak-dose, resulting from supersensitive dopamine receptors in the striatum undergoing pulsatile stimulation.
• Non-motor side effects — nausea, orthostatic hypotension, psychosis (excess dopamine in mesolimbic pathways).

Deep brain stimulation (DBS): High-frequency electrical stimulation ($\sim 130$ Hz) delivered through stereotactically implanted electrodes targeting the subthalamic nucleus (STN) or the internal globus pallidus (GPi). DBS mimics the effect of a lesion (high-frequency stimulation effectively silences or jams the local neural activity) while being reversible and adjustable. DBS of STN or GPi reduces bradykinesia, rigidity, and tremor, allowing reduction of L-DOPA dose and thereby reducing dyskinesias. The mechanism is not fully understood but likely involves disruption of pathological oscillatory activity (exaggerated beta-band oscillations in the STN-GPi circuit that characterise the Parkinsonian state).

Connections Master

This unit extends the gross neuroanatomy of 18.05.01 into the functional specialisation of specific brain regions. The cerebral cortex sections build directly on the Brodmann areas and cortical organisation introduced in the earlier unit. The basal ganglia direct/indirect pathway model connects the nigrostriatal dopamine system to the motor deficits of Parkinson's disease, providing the pharmacological rationale for L-DOPA therapy and the surgical rationale for DBS. The cerebellar error-correction model complements the basal ganglia's movement selection function: the basal ganglia choose which movement to make, and the cerebellum ensures that movement is executed accurately.

The limbic system sections connect to endocrine regulation 18.07.01 through the hypothalamus-amygdala axis (stress response, HPA axis activation). The hippocampal memory system bridges to the cellular mechanisms of synaptic plasticity (LTP, LTD) that are the molecular substrate of all learning. The prefrontal cortex sections on executive function provide the neural basis for the cognitive control processes that modulate attention, emotion, and behaviour throughout the nervous system.

Stroke localisation ties the entire unit together: the clinical syndromes produced by vascular occlusions in specific arterial territories map directly onto the functional anatomy of the cortex, basal ganglia, thalamus, and internal capsule described in this and the preceding units. Clinical neurology is applied functional neuroanatomy.

Historical & philosophical context Master

Korbinian Brodmann's 1909 cytoarchitectonic map ^{[Brodmann 1909]} established that the cerebral cortex is not homogeneous but consists of dozens of distinct areas defined by differences in cell density, laminar thickness, and cell morphology. Brodmann's 52 areas were defined purely from histological stains (Nissl stain for cell bodies) without any knowledge of function. The subsequent mapping of function onto these areas — area 4 for motor, areas 3-1-2 for somatosensory, area 17 for visual, areas 44-45 for speech — was one of the great convergences of structural and functional neuroscience. The fact that a map drawn from cell-staining patterns predicts functional specialisation with remarkable accuracy is strong evidence that cortical function is intimately tied to cortical microcircuit architecture.

Wildler Penfield's intra-operative stimulation mapping (1930s-1950s) provided the first direct evidence for functional localisation in the human brain. By stimulating the exposed cortex of awake patients during epilepsy surgery and recording their reports and observed movements, Penfield and colleagues mapped the motor and sensory homunculi, identified Broca's and Wernicke's areas in vivo, and established the organisation of the temporal lobe for language and memory. The homunculi — distorted human figures with giant hands and lips — became iconic images of brain organisation.

The case of patient H.M. (Henry Molaison, described by Scoville and Milner in 1957) is one of the most influential case studies in neuroscience. Bilateral hippocampal removal produced a profound and selective anterograde amnesia that revealed the double dissociation between declarative memory (lost) and procedural memory (intact: H.M. could learn mirror-drawing and other motor skills without any awareness of having practiced them). Brenda Milner's careful testing of H.M. over decades established the modal model of memory systems: working memory (prefrontal), declarative memory (hippocampus-dependent), procedural memory (basal ganglia/cerebellum), and the distinction between encoding, consolidation, and retrieval.

The discovery of place cells (O'Keefe and Dostrovsky 1971) and grid cells (Hafting et al. 2005) transformed the hippocampus from a generic memory structure into a spatial computation engine with an identifiable neural code. The 2014 Nobel Prize (O'Keefe and the Moser couple) recognised this as one of the most significant advances in systems neuroscience.

Bibliography Master

1. Brodmann, K. — Vergleichende Lokalisationslehre der Grosshirnrinde in ihren Prinzipien dargestellt auf Grund des Zellenbaues. Barth, Leipzig (1909).

2. Penfield, W. & Boldrey, E. — Somatic motor and sensory representation in the cerebral cortex of man as studied by electrical stimulation. Brain 60, 389-443 (1937).

3. Scoville, W. B. & Milner, B. — Loss of recent memory after bilateral hippocampal lesions. J. Neurol. Neurosurg. Psychiatry 20, 11-21 (1957).

4. O'Keefe, J. & Dostrovsky, J. — The hippocampus as a spatial map: preliminary evidence from unit activity in the freely-moving rat. Brain Res. 34, 171-175 (1971).

5. Marr, D. — A theory of cerebellar cortex. J. Physiol. 202, 437-470 (1969).

6. Ito, M. — Cerebellar control of the vestibulo-ocular reflex — around the flocculus hypothesis. Annu. Rev. Neurosci. 5, 275-296 (1982).

7. Hafting, T., Fyhn, M., Molden, S., Moser, M.-B. & Moser, E. I. — Microstructure of a spatial map in the entorhinal cortex. Nature 436, 801-806 (2005).

8. Kandel, E. R., Koester, J. D., Mack, S. H. & Siegelbaum, S. A. — Principles of Neural Science, 6th ed. McGraw-Hill (2021).

9. Sherwood, L. — Human Physiology: From Cells to Systems, 9th ed. Cengage (2016).

10. Silverthorn, D. U. — Human Physiology: An Integrated Approach, 8th ed. Pearson (2019).

11. Schmahmann, J. D. — An emerging concept: the cerebellar contribution to higher function. Arch. Neurol. 48, 1178-1187 (1991).