Final Motor systems

What might be one of the main purpose of the brain? Hint this is a slidedeck based on motor systems

Wolpert argues that we have brains “in order to move

stunning comparative example of this is the sea squirt.

Juvenile sea squirts swim around looking for a comfortable rock to sit on where they can passively extract food from the water for the rest of their adult lives.

Once they settle in, the first thing they do is eat their own brains.

rther, because movement is critical to animals’ survival and reproduction, he argues that perception, attention, learning, memory, etc. generally evolved to support adaptive movemen

What is the motor system good at?

Hierarchically structured movement control

Reflexes are coded in central pattern generators in the spinal cord. Some adaptation based on sensory feedback is also spinally mediated, but much involves the cerebellum. Initiation/gating and goal-related feedback related to action goals involves the basal ganglia.

2. Prediction

Why you can’t tickle yourself Error correction and motor learning

3. Representation

Response representations in motor cortex How you understand other people?

What do central pattern generators do? Where are they located? What is the significance?

Central pattern generators in the spinal cord can generate simple, reflexive actions.

Here, hindlimbs are disconnected from the brain via transection of the spinal cord...

...but sensory feedback about relative position (the force of the ground on the paw, the extension of muscles by stretching out on the treadmill, etc.) is coupled with rhythm generation to produce walking behavior without intervention from above the level of the spinal cord

What kind of activity can be done without a cerebral cortex?

Non-cerebral motor control: the cerebellum and the basal ganglia.

What does coordination from limbs require, which brain area? What is a example of a task that doctors use to test coordination?

Controlled movement generally involves coordination of a large number of muscles and joints.

Touch your nose with one finger, and then, with your finger still pointed, reach out so that you’re pointing straight ahead. Try to be aware of the muscles and joints that are active while you do this.

This coordination seems effortless and “ballistic,” but actually requires considerable sensory feedback from proprioceptive, tactile, and visual sensory information

Damage to the cerebellum due to brain injury or disease reveals a critical role for cerebellar feedback for smooth, volitional movements.

Tremors observed in this example are the result of impaired cerebellar feedback.

How does the basal ganglia inhibit/disinhibit activity?

The basal ganglia (here caudate and SNr) play a direct role in motor initiation/selection.

Activity in the caudate (striatum) inhibits the substantia nigra (pallidum), which in turn tonically inhibits the superior colliculus, a midbrain region involved in eye movements.

This sort of loop (and other, more complex loops involving thalamus and cortex) characterizes the basal ganglia’s role in controlling a wide variety of behavior

What tasks other than motor is the basal ganlgia involved in?

Recall that the basal ganglia are also involved in encoding reward value as well as “habit learning” tasks. (Now it should be clear why motor learning tasks would not be a good test of learning in individuals with Parkinson’s diseas

What kind of feedback does the cerebellum and basal ganglia have?

Feedback is often thought of as functioning like an error signal.

Cerebellum: Is my arm where I expected it to be? If not, adjust...

BG: Did my movement result in expected reward? If not, adjust... But we can’t have an expectation without some kind of prediction.

Why can’t you tickle yourself? Describe the robot arm tickle experiments and how you could increase the tickling experience?

Idea: When you try to tickle yourself, your motor system generates a prediction about the tactile consequences of your movements. Experience provides you with a pretty good model of what it’s going to feel like to touch your own palm, so there’s no discrepancy between predicted and observed stimulation... ...so, no tickliness

We can fix this with a robot arm.

Blakemore et al. got participants to tickle themselves by proxy.

Participants manipulate a tickling rod, and a robot arm holding an identical rod over their other hand generates movement in real time.

At zero delay, it is more or less as if they are still tickling themselves.

Increased delay leads to a greater tickling sensation, so that by 200- 300ms, they are just as tickled as if they were not producing the movement themselves.

Rotating the angle of the robot arm also creates a prediction error, leading to greater tickling sensation... ...and both manipulations increase activity in the cerebellum

Moreover

What part of the brain is muscle memory associated with?

Recall that prediction error is a key component in procedural learning in the basal ganglia/midbrain dopaminergic system.

Similarly, error signals generated by the cerebellum are involved in learning new motor skills.

When people talk about “muscle memory,” they are generally talking about learning involving the cerebellum

Describe the mouse 128 degree experiment. What were the results?

amizu et al. (2000) had people learn to use a “new tool” (their words) in an FMRI experiment. Basically, people had to learn to control a mouse that moved at an oblique angle to their hand movement

Initially, activity is observed throughout the cerebellum, which is thought to reflect the large error signal.

After training, a subset of cerebellar regions remain active, even relative to an “error-equalized” baseline condition (they forced participants to go faster).

This is thought to reflect the acquisition of a new cerebellar circuit for this “tool.

In the joystick experiment where was activity observed?

n a similar experiment, Graydon et al. (2005), trained people to use a joystick that either functioned normally (control) transformed their movements by 90 degrees (motor learning).

Activity in the putamen (basal ganglia) was observed, in addition to activity in the cerebellum during motor learning

A nice demonstration of motor learning is the game QWOP. You are trying to make an avatar run by flexing and adducing his thigh and calf muscles using the Q, W, O, and P keys. It’s hard in part because each key controls both legs, making the controls very counterintuitive.

What do we need the motor cortex for if the basal ganglia and the cerebellum are used for movement?

The basal ganglia are key for selection and initiation of motor outputs, and for learning from reward feedback. The cerebellum is key for integrating online sensory feedback, and for learning new motor skills. So why do we have a big chunk of cortex given over to motor representations

rimary motor cortex has a “direct line” to spinal motor neurons. Modules in motor cortex are organized to represent articulators rather than individual muscles.

This explains the over-representation of the fingers, lips, jaw and tongue -- relative to, say, the trunk, which is much larger, but has little use for fine motor control.

Recall the study by Roitman & Shadlen (2002), in which higher levels of coherence in a random dot motion display were associated with activity encoding related to decision-making in LIP (also FEF, and all the way down in the superior collicus

Describe another random dot display with stimulation in FEF. How do we react when movement is in the opposite direction we expect?

old & Shadlen (2000) combined random dot displays with stimulation in the FEF. The green arrow (middle) indicates the usual, unbiased direction of movement evoked by stimulation, whereas the red and blue arrows show how the direction of movement is influenced by the direction of coherent motion in the dot display. Further, the magnitude of the bias was directly related to the coherence of the motion.

There isn’t just a bias, but a magnitude attached to the bias

What is a mirror neuron?

many neurons in premotor cortex have so-called “mirror” properties.

That is, they fire both during execution of an action, and during observation of the same action.

Further, they seem to encode information relevant to context and intentionality

As a monkey observed someone reaching for something, what happened in mirror neurons? What does the motor activity seem to encode?

Interestingly, motor activity in these cells seems to encode something relatively abstract

Describe how mirror neurons in a monkey react when they see someone pick something up with a tool. How is the activity timed?

In monkeys trained to use tools to pick up food, mirror neurons fire while watching a person pick up food with a tool. Activity is timed to the goal state (grabbing the raisin), not to particular hand movements. (Note that using the “reverse pliers” involves relaxing your grip to tighten the pliers...)

Describe the occluder experiment with mirror neurons

Describe the auditory consequences experiment with mirror neurons

Summary

All of this suggests that mirror neurons are coding something much more abstract than a particular motor plan. They respond to viewing actions performed: - by humans - with tools - behind an occluder - invisibly ...as long as the action has the same goal.

Mirror neurons in humans; where have sensory-motor neurons been studied the most?

Region F5, where these sensory-motor mirror neurons have been studied most, is homologous to Brodmann’s Areas 44/45 (a.k.a. IFG, a.k.a Broca’s area) in humans

No notes slides 36 onwards