Training

Training behaviours coordinate the sequences of randomly targetted trials and marker communication necessary for the back end to train a classifier. The AutomatedTrainingBehaviour and constituent components will serve most use cases.

Custom Sequences

Arbitrary trial behaviour can be implemented by extending the base TrainingBehaviour class, a CoroutineBehaviour. Use of this base class provides interoperability with other behaviour classes.

A motor imagery training behaviour could be as simple as flipping between an “active” and “idle” state every 2 seconds.

Training Flow

Though any matter of sequence may be implemented by a custom training behaviour, the following abstract sequence provides a rough framework.

• Run Training “Round”
  • Choose + Indicate Target
  • Mark Training Epoch or Run Stimulus Trial
• Repeat until Sufficient Training Data
• Indicate Training Completion