Planning Uncertainty

"Planning Uncertainty" by Giuseppe Pellizzer, Brain Sciences Center, Minneapolis Veterans Affairs Medical Center & Dept. of Neuroscience, University of Minnesota

Most experiments about motor control: implemented with equiprobable targets. However, we also know that some motor responses are more likely to be required than others. But the amount of information available is dependent on the context.

[POSTSCRIPT: Sorry, but I had some major problems following this presentation so halfway through I stopped typing regularly, because what had come before was mostly gibberish.]

Hick-Hyman Law: The reaction time is a function of the number of stimuli/responses:

RT = a + b Log_2 (N)

A good description for conditions where the stimulus and response are not very compatible. But it is poorer for desccribing conditions of higher compatibility.


Topic: Effect of directional uncertainty in reaching tasks (high stimulus-response compatibility)

Instructed-delay task: after delay, moving dot to center of other dots (I really didn't quite follow the intricacies. I was distracted by my brownie).

Hypotheses for motor preparation with discrete spatial cues: the system engaged in processing multiple directional info:
1. has a limited processing capacity;
2. can process multiple alternatives simultaneously
3. spends time to initiate a motor response in relation to the amount of processing capacity attributed to the direction of the target.

Capacity-sharing model: RT = a + b(1-1/N)
--pretty good fit of data from instructed-delay task
--RT not related to spatial dispersion of cues
--but some subject populations seem to fit to Hick-Hyman law

Instructed-delay task: continuous cue
--present a range of direction in which the dot can appear, which will be moved to the center.

Instructed-delay task: MONKEYS!!!
--0 cues, 1 cue, 2 cues (45 deg from each other), 2 cues (135 degs), 4 cues (135 deg), 4 cues (270 deg)

We know: many motor cortex neurons are tuned to direction of movement, mapped in orderly fashion.

Q (his): "What determines the level of neuronal activity when there are multiple directional cues?"

4 hypotheses:
1: ? (Most powerful cue, I think)
2: Averaged direction of cues.
3: Sum of neuron's activity associated with each cue
4: Complicated: activity modulated by the average direction with motivational activity affected by dispersion of cues (no, I can't really parse that either)

#4 has highest R^2 when matching data.

Neuronal activity: not dependent solely on cue # or solely on cue range.
--reaction time is inversely correlated with level of neuronal activity (unsurprising)


Conclusions:
--Humans: results suggests that human subjects processed differently discrete and continuous spatial uncertainty
--Monkeys: spatial distribution of cues was an important determinant of neuronal activity and behavior (i.e., reaction time)
--activity of motor & dorsal premotor neurons reflected the spatial uncertainty of the movement to be selected
--monkeys processed similarly discrete and continuous spatial uncertainty