Models of motor control (wk 2)

Question 1

Describe bell-shaped velocity profiles:

Answer 1

-Bell-shaped velocity profiles are one of the most presentative hallmarks of human motor control

Question 2

Describe how EPH plans trajectories:

Answer 2

-Planning the trajectory -> Rotating room experiment showed that EPH is not enough to explain the control of reaching. It suggests the brin does not only specify the EP, but also the entire trajectory (the direct cortical control model)

Question 3

Describe the optimal controls model of movement:
-Includes optimisation

Answer 3

-> Optimal control model assumes that the controller tries to minimise (or maximise) a certain cost (or benefit) produced by the resultant action
-Optimal control is based on the optimisation -> Optimal control models assumes that the brain stores policies (feedback), not the commands (feedforward) (i.e. plan) – optimal feedback control. This control assumes that the brain keeps feedback policies for different motor skills, and having policies rather than plans provides more advantages.
-Optimisation for movement -> Finding optimal trajectory (movement) requires exhaustive search among an infinite number of possible trajectories.

Question 4

Describe the 3 properties to optimise:

Answer 4

• Smoothness (kinematic) -> The brain minimises the jerk of the trajectory (Flash and Hogan 1984). Displacement-> speed -> jerk
• Torques and energy (kinetics) -> The brain minimises the change of torque of the trajectory (Uno et al 1989). Unlike kinematics-based models, this model can predict kinetic effects (inertia, forces, etc)
• Uncertainty -> The brain minimises the uncertainty of reaching (Harris and Wolpert, Nature 1998). It’s considered as a predominant model of the control of reaching.

Question 5

Describe signal dependent noise in the minimum uncertainty model:

Answer 5

-Signal dependent noise (uncertainty) -> Movement causes the noise, whose size depends on the control signal (input to the muscle). Bigger and more abrupt movement causes bigger noise = increased uncertainty. To minimise uncertainty, movements must be controlled smoothly and as slowly as possible.

Question 6

What makes the minimum uncertainty model better than past models?

Answer 6

• It’s a probabilistic model as it accounts for the variability of movement
• Therefore, it provides an inclusive explanation on the empirical laws of movement, such as Fitt’s law and the two-thirds power law

Question 7

Describe optimal feedback control:
-4 bullet points

Answer 7

-Optimal feedback control assumes that the brain keeps feedback policies for motor skills
-Policies rather than plans brings important advantages, such as: no need to store actions individually, no need to plan ahead, robust to the error
-Feedback refers to action generates by a policy relates to a state of the body, instead of a sensory feedback (reflex)
-Properties to optimise – signal dependent noise -> Optimal feedback control well predicts the variabilities of movement with different numbers of via-points