Kayla Degrand

  Faculty of Science and Engineering| University College Maastricht

Less is more: Can a single arousal metric tell us the whole activity of the brain?

Kayla's elevator pitch
Underlying brain processes and fluctuations in brain activity have been a central research interest in Cognitive Sciences. Using electroencephalography (EEG) data, this complex activity can be measured. Since the awareness of an organism impacts brain activity, that arousal state is reflected in EEG signals. Previous studies have shown that cortex activity can be determined using only time measurements of an arousal indicator linked to locomotion in mice specimens. My thesis tested the predictability of human EEG data using temporal eye-tracking. Variational autoencoders were used as my primary machine learning models to estimate brain activity. These prediction models succeeded at recognizing some patterns within the data, providing significant results in regards to brain dynamics.