Motor Memories for Expectation and Uncertainty in the Environment
Metadata
Show full item recordCitation
Hadjiosif, Alkis M. 2015. Motor Memories for Expectation and Uncertainty in the Environment. Doctoral dissertation, Harvard University, Graduate School of Arts & Sciences.Abstract
The motor system endows us with the ability to control movement – an ability that provides us, almost exclusively, with the means through which we interact with our environment. To accomplish this, the motor system continuously maintains, adapts, or forms anew memories that support the execution of a multitude of motor tasks, from the basic but critical like breathing or reaching, to complex and elaborate patterns of motion like those of an NBA player or a master cellist. Here, I use a series of behavioral paradigms to probe the mechanisms behind the formation and retention of motor memories related to motor adaptation. Specifically, I investigate the learning of both internal representations of the expected state of the environment but also representations of the uncertainty associated with that expectation.First, I focus on the interplay between the formation of estimates for the expected environmental dynamics and the formation of estimates about the variability of these dynamics. I show how variability estimates are critical for the control of the safety margin associated with grip forces, and that grip force control is, in fact, three times more sensitive to changes in variability – specifically, the standard deviation, σ – than the expected value of environmental dynamics. This ratio of sensitivities effectively acts to provide a 3-σ confidence level (>99%) against slip. A consequence of heightened sensitivity of grip forces to uncertainty that we observe, is that grip forces strongly increase upon a change in environmental dynamics, regardless of whether that change increased or decreased the strength of the load. This in turn leads to bizarre, asymmetric grip force learning curves to step increases vs. step decreases in dynamics, explaining previous studies where grip force adaptation appeared to outpace manipulatory force adaptation. I proceed to show that this variability-driven safety margin control is based on a low-order variability metric that is robust to outliers in the distribution of environmental dynamics.
Second, I examine the temporal stability of motor adaptation. Using a visuomotor adaptation task, I investigate the speed and breadth of decay due to the passage of time, and find that adaptation consists of distinct temporally-labile and temporally-stable components. Whereas temporally-labile adaptation decays within the timecourse of only a few tens of seconds, the remaining temporally-stable adaptation persists well beyond that timeframe. Moreover, this temporally-stable adaptation predicts and acts as a gateway towards the long-term retention of the trained adaptation. Comparison of the temporal stability of motor adaptation formed from one training session to the next reveals wide-ranging differences that are not only driven by differences across individuals, but, remarkably, by considerable differences across different sessions for the same individual.
Based on a brain stimulation study and the reanalysis of previous studies on cerebellar patients, I proceed to identify the primary motor cortex and the posterior parietal cortex as likely parts of a network involved with temporally-labile adaptation, and the cerebellum as a part of a network involved with temporally-stable adaptation. A different study revealed that, in spite of the fact that temporally-stable adaptation is remarkably persistent even after prolonged active washout, savings – the phenomenon of faster relearning of a previously learned adaptation – is not driven by the reemergence of this strong, stable memory, but, instead, by an increased propensity of the temporally-labile component to relearn the task at hand.
Third, I examine how the rate of adaptation to changes in the environment is modulated by the statistics of that environment. Specifically, I show how the rate at which the motor system adapts to changes in the environment is primarily determined not by the degree to which environment change occurs – i.e. environmental variability – but by the degree to which the changes that do occur persist from one movement to the next – i.e. the trial-to-trial consistency of the environment. Finally, I find that repetition of the same learning stimulus can greatly potentiate the effect of consistency, although, unlike consistency, repetition alone does not increase adaptation rate.
Terms of Use
This article is made available under the terms and conditions applicable to Other Posted Material, as set forth at http://nrs.harvard.edu/urn-3:HUL.InstRepos:dash.current.terms-of-use#LAACitable link to this page
http://nrs.harvard.edu/urn-3:HUL.InstRepos:17467305
Collections
- FAS Theses and Dissertations [6136]
Contact administrator regarding this item (to report mistakes or request changes)