Physiomotor responses

Let’s analyse the neurophysiological and motor data as “emotional responses” rather than “EEG artefacts”.

Possible response types

• vEOG: blink rates
• hEOG: Mouse-cursor-locked horizontal EOG? meaningless?
• facial EMG: smiling and frowning?
• SCR: event-related phasic responses
• RESP: breathing rate variability?
• PULSE: heart rate variability?
• ACC: frequency-tagging: magnitude; or phase-locking.

TODOs

• Neurokit2 preprocessing
• Motor-beat coherence with tempoCNN (but not too convincing)
• BEAT-THIS! (ismir-2024 SOTA) for beat tracking
• Jerk beat contrast

Motor-beat coherence

• Toiviainen & Carlson (2021) proposed a wavelet tensor decomposition to analyse motion-capture data acquired during spontaneous dance to musical excerpts. [BUT DO I NEED IT?]

• Ito et al. (2022) showed spontaneous beat synchronization in rats using “jerk (dA/dt) beat contrast”. This time-based methods requires a beat annotation. (but frequency tagging also needs a tempo estimation, and the assumption of constant tempo). [HMMPH!]