Research interests
I want to understand how useful information can be recovered from brain activity, especially when the signals are noisy and change over time.
EEG and neural decoding
My starting point is automated sleep staging from EEG: learning to associate a segment of a recording with a sleep-stage label. I completed an internal study using the Sleep-EDF dataset family on PhysioNet, with person-separated development and reserved-person evaluation. Publication and deployment remain pending.
A question I want to investigate is how a model's performance changes across people and recordings, rather than only across segments from familiar recordings. Projects
Brain–computer interfaces
I am interested in decoding movement-related activity. An initial direction is motor-imagery classification: using recordings collected during an instructed imagery task to classify task labels. The PhysioNet EEG Motor Movement/Imagery dataset is a candidate resource. I have not yet selected an experiment or implemented a decoder.
Neural dynamics and spiking models
I am learning how simplified neuron models describe changes in membrane potential and the timing of spikes. The interactive LIF explorable is part of that learning. The integrate-and-fire chapter in Neuronal Dynamics gives a mathematical starting point: simplified voltage dynamics and a threshold/reset rule. I have not built an SNN research project or measured energy savings.
Neuro-AI
I am interested in the relationship between neural computation and machine learning: what can biological systems suggest about temporal representations, and what can computational models help us explain? At present, this is a learning and research direction.