I collected the data from the first participant for my pegboard task this week. It’s a simple pilot study where we aim to investigate how sensory uncertainty affects motor learning during a fine motor control task. The participant simply has to fill holes on a pegboard with pegs, but one group wears a blindfold to eliminate vision and medical style nitrile gloves to reduce tactile input slightly, and the other group has full vision and full tactile input available. They train for one day like this, and then on the second day they both have the same training – blindfold but no gloves. We intersperse minute long tests to see how many peg ‘towers’ they can build with and without vision, and we can see how these values converge or disperse over the course of the two days as well as examine the learning curves.
You might expect the group without the gloves and with full vision available to not just perform better on the first training day, but also to continue with higher performance on the second day too. They have of course had more sensory information available to them to better learn the task. The task conditions on the second day though is new to both groups – same pegs and pegboard, but new combo of blindfolds without gloves.
The other possibility is that the group with reduced sensory information (gloves and blindfolds) alters the way they control movement. We’re hoping to shift the way this group controls movement, relying on prediction more so than external sensory information. This shift toward predictive control occurs when we move in an uncertain or noisy environment and can no longer rely so heavily on sensory feedback we get while moving. Combining (1) predictive information of what we expect the sensory consequences of our movements to be based on previous movements with (2) actual and current sensory input is a more efficient way to control movement than just using one or the other. What we don’t know, is how increasing predictive control might affect motor learning, and whether by increasing it acutely only has a transient effect, or actually lasts.
A recent paper by Camilla Pierella and colleagues attempted to get at the fundamentals of motor learning by messing with the internal models (forward and inverse). In their mathematical model, errors and inputs drive the dynamic learning process. This pegboard study has one condition that aims to increase the weighting of the predictive signal which might resultantly increase error signal when normal tactile input is restored, and the other condition maximises task-relevant sensory input. Is either one better for learning than the other? If they’re the same, what does that tell us about the role of predictive control in motor learning? Will the effect of vision override any effects of altering tactile input slightly?
We aim to collect data from 10 participants in each group in the next month or so and have a look at the data before deciding to go any further with it. Whilst it’s not one of the main studies that will form their own chapters in my thesis, there are some really intriguing questions behind it, and I’m eager to find out what happens, if anything!