The basal ganglia: a functional approach
“But, what is it for?” Finding out with computational neuroscience
1 Phenomenology
● Damage to BG is main cause of: ■ Parkinson’s disease PD video ✴ Bradykinesia, akiensia, tremor
■ Huntington’s disease
✴ Involuntary movements (when slow and extended = chorea)
● Damage to BG implicated in:
■ ADHD
✴ Attention deficit hyperactivity disorder
✴ Complex condition with positive symptoms (e.g. delusions) and negative symptoms (e.g. lack of affect) 2 Rationale for tackling the problem
● If brain area/system X is damaged, what are the consequences?
● Strategy: find out the main function of X.
● Then, understanding of the phenomena will be unified: each deficit is an aspect of loss of the main function of X.
● So, what do the basal ganglia do….?
3 Action selection
Predisposing conditions (sensory input, cognitive state, homeostasis, etc)
Energy balance Fluid balance Threat (A3: (A1: feeding) (A2: drinking) escape)
Motor resources
Behavioural output 4 (feeding) Cognitive example
● Students in this lecture
■ ‘shall I pay attention or think about plans for this evening’? [‘action’ of defining cognitive state]
■ ‘pay attention or act on my being hungry’
■ but then fire alarm goes off….
● Salience issues (again)
● ‘lower’ and ‘higher’ level processing
■ social protocols and long term planning versus biological drives and homeostasis
■ Probably mediated by cortical and sub-cortical
5 centres respectively Mediating competition between actions
Central switch has a ‘wiring’ advantage over the ‘distributed’ alternative
A1 A1
A1 A1 A1 S A1
A1 A1
Distributed (all-to-all) Central switch connectivity. Number of connections ~ Number of connections ~ N 6 N2 Hypothesis for basal ganglia function
● basal ganglia is the central ‘switch’ in the vertebrate brain that enables ‘action’ selection
■ Here ‘action’ may mean ‘internalised behaviour’ defined by cognitive state ● Therefore disorders of the basal ganglia are disorders of action selection
■ E.g. PD, HD = inability to select motor actions appropriately
■ E.g. ADHD, Schiz. = inability to suppress unwanted actions or cognitive states 7 Redgrave ,et al. 1999 Evidence for this hypothesis: 1
If the action selection hypothesis is true then we require that …
1. The BG must take input from a wide variety of brain areas • i.e. putative ‘command centres’ for ‘action requests’ • These should be cortical and subcortical 2. There must be some mechanism by which the BG can switch on selected motor resources 3. There must be evidence of a flow of information between command centres, BG, and associated motor centres 8 Evidence for this hypothesis: 2
If the action selection hypothesis is true then we require that …
4. The BG must understand be able to extract the ‘salience’ of each request • Assuming that this is done in BG and is not sent separately (otherwise we have a repeat of the ‘wiring argument’ ) 5. There must be evidence of neuronal selection mechanisms in BG
Subsequent slides will be related back to these points by labels of the form E# 9 Basal ganglia input E1 globus pallidus The main input nucleus is putamen neostriatum (or sometimes simply the striatum) composed of the putamen and caudate nuclei.
The striatum is by far the largest nucleus in the basal ganglia and receives input from all over the brain including: all of cerebral cortex (except primary sensory areas) and subcortical nuclei such as the superior colliculus (via the intralaminar thalamus) head of caudate tail of Also shown is the globus caudate pallidus (see next slide for 10 details) Disinhibition gating hypothesis 1: basic mechanism E2 Cortex
Thalamus striatum striatum BG Tonic inhibition GPi/SNr GPi/SNr
subcortical motor nuclei Output nuclei of the BG GPi = globus pallidus internal segment SNr = substantia nigra pars reticulata
11 Alexander,et al. 1990 Disinhibition gating hypothesis 2: data with glutamate injections in striatum E2
VM = ventromedial thalamus
SC =superior colliculus (midbrain sensory and striatum motor motor nucleus)
VM
Alexander,et al. 1990 SC 12 SNr Representing discrete actions 1: cortico- basal ganglia loops E3 Five main loops according Cortex to primary cortical input area
‘Motor’ (supplementary motor area – SMA) Thalamus BG ‘Oculomotor’ (frontal eye fields – FEF)
‘Prefrontal 1’ (dorsolateral prefrontal cortex – dlPFC) Brainstem Motor nuclei ‘Prefrontal 2’ (lateral orbitofrontal prefrontal cortex – LOF)
13 ‘limbic’ (anterior cingulate area – ACA) Representing discrete actions 2: topography in BG E3
Different body parts represented by different parts of BG and related structures
We conclude (from this slide and the last) that BG represents actions in discrete channels
14 Representing discrete actions 1: sub- cortical-basal ganglia loops E3
Sub-cortical structures Sub-cortical structures include: Superior colliculus (SC)
Intra-laminar thalamus Pedunculopontine tegmental area (or simply peduncularpontine nucleus - PPN)
BG Periaqueductal gray (PAG)
Brainstem 15 Motor nuclei Disinhibition gating and channels
Predisposing conditions E2 & E3
Ctx1: action1 Ctx2: action2
Thalamus Thalamus BG BG
Motor resources Motor resources Probably some shared for action 1 for action 2 resources
16 Interlude – the anatomy globus pallidus Coronal putamen section
head of caudate tail of caudate
output nuclei
17 Salience extraction in striatum Medium spiny neurons E4
18 Spines up close…
DS
Sp
AT
19 textbook Paul Bolam’s lab Selection mechanisms in BG 1: basic idea
E5
Cortical inputs ‘striatum’ Note: ‘striatum’ is both excitatory and inhibitory Resolved later...
Output (GPi/SNr)
20 inhibition excitation Model neuron Selection mechanisms in BG 2: realistic Instantiation in basal ganglia E5 inputs
input - + Striatum STN striatum STN output
‘output layer’ Diffuse projection
21 Assumes diffuse projections from STN = subthalamic nucleus Mink and Thach,1993 Focussed and diffuse connectivity
Medium spiny neuron
STN neuron 22 Striatal structure and BG architecture
Mainly D1 dopamine receptors
Mainly D2 dopamine receptors
NB – red/blue do NOT signify excitation/inhibition here
23 Instantiation in basal ganglia: 2 E5
Cortex/ Cortex/ thalamus thalamus
Striatum (D1) STN Striatum (D2) STN Selection ? pathway EP/SNr GPe
BG output Function?
24 New functional architecture: selection and control pathways
Cortex/thalamus Gurney et al, 2001
E5 Striatum (D1) STN Striatum (D2)
Interpret GP EP/SNr GP efferents as control signals for modulating selection pathway
25 Selection pathway Control pathway Simulation - basic selection dynamics
salience
26 EP/SNr output Comparison with prevailing model: ‘Direct’ and ‘Indirect’ pathways
Cortex/thalamus
Striatum (D2) Striatum (D1)
GP
STN
EP/SNr
Selection Direct Control • Never instantiated 27 quantitatively Indirect
• No global selection Summary
● Action selection is a key computational problem for all animals. ● Central switching appears to be an optimal solution ● The BG appears well placed to do this in vertebrates because
■ It receives widespread input (cortical and subcortical)
■ Tonic inhibition and its selective release appear to be a mechanism for gating motor programmes
■ Functional loops and topography support the idea of action ‘channels’.
■ Medium spiny neurons could act to extract salience
■ There is a multitude of selection mechanisms in BG, including that supported by the innervation of SNr/GPi from striatum and STN. ● Under our hypothesis, disorders of the basal ganglia 28 are understood as disorders of action selection Suggested reading
● The chapter in Squire et a (Fundamental Neuroscience)
● The review by Mink (1996)
● Our paper outlining the action selection hypothesis (Redgrave et al 1999)
29 References
Alexander, G. E. and M.D., C. (1990): Functional architecture of basal ganglia circuits: neural substrates of parallel processing. Trends in Neuroscience 13, 266-271.
Redgrave, P., Prescott, T. J. and Gurney, K. (1999): The basal ganglia: a vertebrate solution to the selection problem? Neuroscience 89, 1009-1023.
Mink, J. W. and Thach, W. T. (1993): Basal ganglia Intrinsic circuits and their role in behavior. Current Opinion in Neurobiology 3, 950-957.
Gurney, K., Prescott, T. J. and Redgrave, P. (2001): A computational model of action selection in the basal ganglia I: A new functional anatomy. Biological Cybernetics 84, 401-410.
Gurney, K., Prescott, T. J. and Redgrave, P. (2001): A computational model of action selection in the basal ganglia II: Analysis and simulation of behaviour. Biological Cybernetics 84, 411-423.
Mink, J. W. (1996): The basal ganglia: Focused selection and inhibition of competing motor programs. Progress in Neurobiology 50, 381-425.
Squire, L. R., M. J., Bloom, McConnell, S.K., Roberts, J. L. Spitzer, N.C. and Zigmond,, F. E. (2003): Fundamental Neuroscience, Academic Press. Chapter 31 30