Neuroanatomy & the Basal Ganglia

1. The Substantia Nigra pars compacta

The substantia nigra — literally “black substance” — is a midbrain nucleus in the ventral tegmental area, immediately dorsal to the cerebral peduncle. It owes its colour to neuromelanin, a cytoplasmic pigment generated as a by-product of dopamine and noradrenaline oxidation. Anatomists since Sömmerring (1791) have distinguished two cytoarchitectonic subdivisions:

• Pars compacta (SNc) — densely packed pigmented dopaminergic neurons (Dahlström and Fuxe’s “A9” group, 1964); ~400,000–500,000 neurons per side in healthy adult human.
• Pars reticulata (SNr) — GABAergic projection neurons that share embryological origin and circuit function with the internal globus pallidus (GPi); the major output nucleus of the basal ganglia.

SNc neurons are arranged in clusters called “nigrosomes” (Damier et al., Brain 1999) on the basis of calbindin-poor / calbindin-rich matrix architecture. Five nigrosomes (N1–N5) lie within the calbindin-rich matrix. Nigrosome 1 (N1) in the dorsal tier is the most vulnerable in PD — it loses ~98% of its neurons by end-stage disease — while ventral tier and matrix neurons are relatively spared. The loss of N1 produces the diagnostic swallow-tail sign on susceptibility- weighted MRI (covered in Part VI).

Pathologically, PD shows ~50–60% SNc dopaminergic neuron lossby the time of clinical diagnosis — this is the threshold below which striatal dopamine compensation fails and motor symptoms become unmasked. Fearnley and Lees (Brain 1991) calibrated this curve: the rate of nigral neuron loss is ~5–10% per year early in disease, slowing to ~1–2%/year in late disease, suggesting an exponential-decay biology.

2. Dopamine Synthesis & Tyrosine Hydroxylase

The dopaminergic phenotype of SNc neurons is defined by the presence of three biosynthetic enzymes:

• Tyrosine hydroxylase (TH) — the rate-limiting step: tyrosine → L-DOPA. Requires tetrahydrobiopterin (BH4), oxygen, and iron. Regulated by feedback inhibition from intracellular catecholamines and by Ser-40 phosphorylation by PKA. Loss of TH activity precedes neuronal death and is itself an early marker of PD.
• Aromatic L-amino acid decarboxylase (AADC, DDC) — L-DOPA → dopamine. Pyridoxal-phosphate (B6) dependent. Also expressed by serotonergic neurons.
• Dopamine β-hydroxylase — converts dopamine to noradrenaline; not expressed in SNc, so SNc neurons stop at dopamine. Expressed in locus coeruleus.

Once synthesised, dopamine is concentrated into synaptic vesicles by VMAT2 (vesicular monoamine transporter 2), the target of reserpine and tetrabenazine. Free cytoplasmic dopamine is cytotoxic — it auto-oxidises to dopamine quinone and reactive oxygen species — so robust VMAT2 sequestration is itself neuroprotective. After release, dopamine is cleared by DAT (dopamine transporter, SLC6A3) and degraded by MAO-A/B and COMT, generating DOPAC and homovanillic acid.

Therapeutically, every step of this pathway has been pharmacologically exploited (Part VII): L-DOPA bypasses TH, carbidopa blocks peripheral AADC, MAO-B inhibitors prolong synaptic dopamine, COMT inhibitors prevent peripheral L-DOPA breakdown, DAT imaging (DaTscan) is the diagnostic tool, and AAV-AADC gene therapy aims at restoring the second enzyme to a degenerated putamen.

3. The Nigrostriatal Pathway

SNc neurons project massive, profusely arborised axons through the medial forebrain bundle to the dorsal striatum — predominantly the putamen (motor striatum) and to a lesser extent the dorsal caudate (associative striatum). A single SNc neuron may give rise to ~150,000–500,000 striatal varicosities and innervate ~5% of the entire striatum (Matsuda et al., J Neurosci 2009). This vast arborisation is a recurring theme in selective vulnerability: the proteomic and bioenergetic burden of maintaining such an axonal network may itself be a key determinant of PD-related cell death.

Striatal dopamine acts on two principal classes of medium spiny neurons (MSNs), defined by which of the dopamine receptor families they express — setting up the direct/indirect pathway architecture below.

4. Architecture of the Basal Ganglia

The basal ganglia are a set of subcortical nuclei that, together with the thalamus, implement a re-entrant cortico-basal-thalamo-cortical loop. The relevant players for movement control:

• Striatum = caudate + putamen + nucleus accumbens. ~95% of striatal neurons are GABAergic medium spiny neurons (MSNs); ~5% are interneurons (cholinergic giant aspiny “TANs”, parvalbumin fast-spiking).
• Globus pallidus externa (GPe) — GABAergic; the relay nucleus of the indirect pathway.
• Globus pallidus interna (GPi) — GABAergic; together with SNr, the major output nucleus of the basal ganglia.
• Subthalamic nucleus (STN) — the only glutamatergic node within the BG; gateway of the indirect and hyperdirect pathways; the most successful DBS target in PD.
• Substantia nigra pars reticulata (SNr) — GABAergic output to thalamus and superior colliculus; functionally the GPi for the head/eye.
• Substantia nigra pars compacta (SNc) — the modulator: dopaminergic input to striatum.
• Pedunculopontine nucleus (PPN) — brainstem cholinergic/glutamatergic node coordinating gait and posture; emerging DBS target for freezing of gait.

MSNs come in two flavours, defined by dopamine receptor and projection target:

Dopamine excites D1-MSNs (Gαs → PKA → phosphorylation of DARPP-32 → enhanced excitability) and inhibits D2-MSNs (Gαi/o → reduced cAMP). The sign asymmetry at the receptor level is the cornerstone of the direct/indirect model.

5. The Direct and Indirect Pathways

The classical model of basal ganglia function was articulated independently by Roger Albin, Anne Young and John Penney (Trends Neurosci 1989) and by Garrett Alexander and Mahlon DeLong (Trends Neurosci 1990; J Neurophysiol 1990). It remains the most clinically useful framework in movement disorders, even after thirty years of caveats.

Direct pathway (Go)

D1-MSNs project monosynaptically to GPi/SNr and inhibit them. GPi tonically inhibits the ventrolateral and ventral-anterior thalamus, which in turn excites motor cortex. Direct pathway activation thus dis-inhibits the thalamus → promotes movement.

Indirect pathway (No-Go)

D2-MSNs project to GPe and inhibit it. GPe normally inhibits STN, so removing GPe tone excites STN, which then excites GPi, which more strongly inhibits the thalamus. Indirect pathway activation suppresses competing movements.

Hyperdirect pathway

Cortex projects monosynaptically to STN, bypassing striatum. Provides a fast global “stop” signal (Nambu et al., Neurosci Res 2002; Aron and Poldrack, J Neurosci 2006). The hyperdirect pathway is increasingly implicated in PD freezing of gait and is a candidate explanation for STN-DBS efficacy.

Net dopamine effect on movement: facilitatory. Loss of dopamine ⇒ D1-MSN under-active + D2-MSN over-active ⇒ STN over-active ⇒ GPi over-active ⇒ thalamic over-inhibition ⇒ bradykinesia.

6. The Model of Bradykinesia

The Albin–DeLong model makes specific, testable predictions about pathological firing in PD — many of which have been confirmed in MPTP-monkey electrophysiology and in human DBS recordings:

• Increased GPi firing rate — from ~70 Hz baseline to ~90–100 Hz in MPTP-monkey (Filion & Tremblay, Brain Res 1991).
• Increased burst-firing in STN and GPi.
• Pathological β-band oscillations (13–30 Hz) in STN, GPi, and motor cortex — the LFP signature recorded by every modern DBS electrode (Brown et al., J Neurosci 2001). Levodopa and DBS both suppress β oscillations as they alleviate bradykinesia.
• Loss of receptive-field specificity in GPi neurons — loss of motor segregation.
• L-DOPA-induced dyskinesias — conversely associated with γ-band (60–90 Hz) oscillations, suggesting a cross-frequency switch.

The therapeutic logic of DBS follows directly: high-frequency (~130–180 Hz) stimulation of STN or GPi disrupts the pathological oscillation by overwriting it with regular antidromic activation, mimicking a functional lesion. The model also predicted (and DeLong’s lab confirmed in MPTP-monkey, Bergman et al., Science 1990) that STN lesion reverses parkinsonism — the experiment that opened the door to STN-DBS in humans (Benabid 1994; Limousin 1995).

7. The Other Dopamine Systems — A8, A10, Hypothalamic

SNc (A9) is not the only dopamine system in the brain, and not all dopamine systems are equally vulnerable in PD. The Dahlström–Fuxe map (Acta Physiol Scand 1964) named four CNS dopaminergic groups by their relative vulnerability:

• A8 (retrorubral field) — small midbrain group; relatively spared in PD.
• A9 (SNc) — nigrostriatal pathway to dorsal striatum; most vulnerable.
• A10 (ventral tegmental area, VTA) — mesolimbic / mesocortical pathway to nucleus accumbens, ventral striatum, prefrontal cortex; partially affected in PD (~30–40% loss). The substrate of dopaminergic agonist-induced impulse-control disorders.
• A11–A15 (hypothalamic) — tuberoinfundibular and incertohypothalamic; relatively spared.
• A16 (olfactory bulb), A17 (retina) — spared.

The selective vulnerability of A9 over A10 is one of PD’s defining puzzles and is partially explained by SNc-specific properties: pacemaker firing dependent on Cav1.3 L-type Ca²⁺ channels (Chan et al., Nature 2007), enormous axonal arborisation, neuromelanin accumulation, mitochondrial complex-I sensitivity to MPP⁺, and high baseline oxidative load. The isradipine hypothesis — that dampening Cav1.3 might protect SNc neurons — was tested in the STEADY-PD III trial (JAMA Neurol 2020) and found negative, but the underlying biology remains relevant.

8. MPTP and the Langston Story

In July 1982, William Langston (then at the Santa Clara Valley Medical Center) admitted a 42-year-old heroin user with profound bradykinesia, rigidity and tremor of one week’s duration — full-blown parkinsonism in a patient four decades too young (Langston et al., Science 1983). Five more young addicts followed, all of whom had injected a designer opioid (“new heroin”) synthesised illicitly in San José. The contaminant was identified as 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP), formed as a synthetic by-product of MPPP (a meperidine analogue).

MPTP itself is lipophilic and crosses the blood-brain barrier readily. Inside astrocytes, MAO-B oxidises it to MPP⁺, a charged pyridinium ion taken up selectively by DAT into dopaminergic neurons, where it concentrates in mitochondria and inhibits complex I (NADH-ubiquinone oxidoreductase)of the electron transport chain. The result: ATP collapse, oxidative stress, and selective SNc cell death within days.

The MPTP story also re-opened the environmental-cause debate: agricultural pesticide exposure (paraquat is structurally similar to MPP⁺; rotenone is a complex-I inhibitor sold as an organic insecticide) is associated with elevated PD risk in multiple cohorts (Tanner et al., Environ Health Perspect 2011; the AHS cohort).

9. Why are SNc Neurons so Vulnerable?

Selective vulnerability is the central puzzle of every neurodegenerative disease. For SNc, multiple converging features make it perhaps the most metabolically precarious cell type in the brain (Sulzer, Trends Neurosci 2007; Surmeier et al., Nat Rev Neurosci 2017):

• Massive axonal arborisation — ~5% of striatum innervated per neuron; an estimated 4× the bioenergetic cost of typical CNS neurons.
• Pacemaker firing — SNc neurons fire spontaneously at 2–10 Hz, driven by L-type Cav1.3 calcium channels (rather than the more energy-efficient Na⁺-mediated pacemaking of A10/VTA neurons). The Ca²⁺ load demands constant ATP-dependent extrusion and mitochondrial buffering.
• Cytoplasmic dopamine — auto-oxidation of dopamine generates DOPAL (the toxic aldehyde of dopamine), dopamine quinone, and ROS; insufficient VMAT2 sequestration is itself toxic.
• Neuromelanin accumulation — iron-binding pigment that, late in life, may shift from chelator to oxidant.
• Complex-I sensitivity — SNc mitochondria show baseline ~30% reduction of complex-I activity in PD; MPTP and rotenone exploit this.
• α-synuclein abundance — SNc neurons express α-synuclein at high levels (it is presynaptic and regulates vesicle dynamics); aggregation prone in this milieu.
• Lysosomal stress — the mountain of α-synuclein, dopamine quinone adducts, and damaged mitochondria places extreme demand on autophagy/mitophagy — the system implicated genetically in PD via PRKN, PINK1, GBA, ATP13A2, and LRRK2.

These features set up the molecular machinery of the next module: Part III — α-Synuclein and Lewy Bodies, where we follow a 140-amino-acid intrinsically disordered protein from synaptic terminal to fibril cryo-EM and into the prion-like spread hypothesis. The genetic landscape (Part IV) maps almost one-to-one onto these vulnerability mechanisms.