SfN abstract

General hypotheses

1. uncontrolled (network constraints) minimum intervention: todorov 2002, liu & todorov 2007, scott 2004 or so
2. minimum energy (network + energetic constraints)
3. task-driven (network + task constraints) (constrained by sensory/motor covariates) habitual: derugy & loeb jneurosci "habitual rather than optimal control"

Sketch of argument

no reason that null should depend on motor or sensory parts of the task

we know that not all activity is possible, that predictions must be possible in the network. other than that, is it

1. no further constraints
2. energetic constraints, or
3. task-dependent

task demands vs. behavioral output (cursor-target angle vs. row space)

Words not to mention in results (maybe in methods):

• factor analysis
• row space
• cursor-target angle
• hypothesis names (stick with the classes)

BCY presentation

Outline:

0. motivation
1. hypotheses to cover: minimum baseline unconstrained habitual cloud pruning
2. how I evaluate them
3. results: representative day