Bio

Trish Shewokis is a Professor at Drexel University with appointments in the Nutrition Sciences Department in the College of Nursing and Health Professions and the School of Biomedical Engineering, Science and Health Systems. Part of her responsibilities at Drexel is to direct research at the Cognitive Neuroengineering Experimental Quantitative Research (CONQUER) Collab0rative.

By training, she is a movement scientist and applied statistician whose research focuses on integrating the brain in the loop when learning multiple tasks—the primary neuroimaging modality of her work is functional near infrared spectroscopy.

Her hobbies include reading, cooking, woodworking, walking, and swimming. Her recent interests include developing expertise in digital photography.

CONQUER Collab0rative: http://www.biomed.drexel.edu/fnir/CONQUER/Welcome.html

JOVE tutorial on fNIR: http://www.jove.com/video/3443/using-mazesuite-functional-near-infrared-spectroscopy-to-study

Project: Assessing and Visualizing Learning of a Bimanual Coordination Task

The role of practice is crucial in the skill acquisition process and for assessments of learning. We know that learning is inferred from changes in behavioral performance across time given that these changes are consistent and persistent. However, assessments of learning are best exemplified by retention (i.e., memory) and transfer (i.e., generalizability) tests, while it is important to first determine differences across acquisition trial blocks as an initial indicator of learning. In addition to testing the different phases of learning, namely, acquisition, retention, and/or transfer, in many instances, individualized learning curves are evaluated for assessment and visualization of within individual skill acquisition processes.

Important for our understanding of learning processes, Speelman and Kirsner (2005) note that learning curves are individualized based on the performers’ previous experiences, and learning a new task is practice of previously acquired skills within a new context.

The aim of this project is to use Mathematica to assess and visualize individual learning curves after 500 practice trials of a novel fine motor coordination task.

I envision the steps in the process to include: 1) Use Mathematica with a controller state to use an Xbox controller interface device; 2) Check a basic two-hand fine motor task created by Christopher Wolfram, then modify this task; 3) Collect data on several subjects with varying degrees of experience with interface controllers; 4) Create appropriate performance metrics and plot learning curves of performance metrics across trials (dependent measures: accuracy, constant error, variable error, and kinematic measures of displacement, velocity, acceleration, and jerk profiles); 5) Assess models of learning the tasks; 6) Include various graphics and plots that will illustrate the various learning strategies exhibited by the performers.

Future work would address two questions: 1) For implicit learning, can individuals predict the future positions of a bimanual coordination tracking task?; and 2) Given a random task like a cellular automaton, can individuals learn this task?

References

[1] C. Speelman, K. Kirsner, Beyond the Learning Curve: The Construction of Mind, London: Oxford University Press, 2005.

[2] C. Wolfram, “Reflex-Testing Notebook for Xbox Controller-Mathematica,” 2013.

Acknowledgements:

Thanks to the comments, insights, and suggestions for this work by Stephen Wolfram, Vitaliy Kaurov, Christopher Wolfram, and Todd Rowland.

Favorite Four-Color, Four State Turing Machine

Rule 14325448312827081304005