Skip to content

A mobile app for research into uncontrolled manifold analysis.

License

Notifications You must be signed in to change notification settings

OlafHaag/NeuroPsyResearchApp

Repository files navigation

NeuroPsy Research App

License: MIT

Get it on Google Play

This app should serve as a starting point for neuropsychological research on mobile devices.
The necessity for conducting research outside of a laboratory became more paramount during the COVID-19 pandemic when face-to-face contact was restricted as a safety measure.

Background

This app demonstrates a toy example for the uncontrolled manifold (UCM) hypothesis (Scholz & Schöner, 1999). In its basic form there is one performance criterion for a task, but two degrees of freedom (df) to fulfill that goal.

target = x1 + x2 | target = const
x1 and x2 are two uncoupled state variables.

This creates a subspace of df configurations that is redundant for successful task performance.

Final state variability in df configurations should be elongated along the task-irrelevant direction and be less in the task-relevant direction orthogonal to it.

This is a toy example as it doesn't really explore the variability in the underlying biomechanical system and only captures results of the control signals. I assume, there's no measurable effect of control-dependent noise either.

Since optimal feedback control is related to UCM theory in the sense that the optimal control law may not act along certain dimensions (the UCM), one prediction made for optimal feedback control in the presence of control-dependent noise is that if more control is asserted in the redundant direction (e.g. no optimal control law) the reduced variance in that direction comes at the expense of increased variance in the task-relevant direction (Todorov, 2004).

To demonstrate one possible operationalization for testing that prediction, there's one task condition that tries to limit the variability compared to the redundant direction of the unconstrained task by introducing a second performance criterion that relies on one of the degrees of freedom. As mentioned, due to suppositional absence of measurable effects of control-dependent noise, no increased error in the first performance criterion is expected.

References

John P. Scholz; Gregor Schöner (1999). "The uncontrolled manifold concept: identifying control variables for a functional task". Experimental Brain Research. 126 (3): 289–306.

Todorov, Emmanuel (2004). "Optimality principles in sensorimotor control". Nature Neuroscience. 7 (9): 907–915.

Installation

Translation

  • Use PoEdit to extract strings wrapped in _("...") function calls.
  • In PoEdit add a new extractor for the kivy language files.
  • To export the Terms of Use and Privacy Policy to markdown files for all translated languages, run the terms.py and privacypolicy.py files respectively.

Legal attribution

Google Play and the Google Play logo are trademarks of Google LLC.