I investigate the laminar-specific responses of the dorsolateral prefrontal cortex during the manipulation of working memory load, revealing that superficial layers exhibit stronger activation to higher load during the delay and retrieval periods and shows a dynamic recoding of this load signal across the trial. 

Links: paper, bioRvix, github, tweeprint

Using subspace temporal cross-decoding and subspace analyses, I show that dynamic coding and temporally stable coding spaces facilitate the concurrent maintenance of perception and visual working memory across the visual cortex. 

Links: paper, bioRvix, github

I train RNNs on a delay-matched-to-sample working memory task that contains task-irrelevant distractors and investigate the subspaces in which memories and distractors are encoded. 

Links: github

We conducted a replication study using automated analysis and found only partial support for previously reported layer-specific activity in the human dlPFC during working memory, highlighting ongoing uncertainties and the need for more reproducible methods in layer fMRI research.

Links: preprint

I use simulations to show that vascular draining distorts layer-specific decoding in GE-BOLD fMRI, but that signal deconvolution can improve specificity—though its effectiveness depends on the layer(s) of origin of the signal.

Links: bioRvix, github, Zenodo

I use high-resolution resting-state fMRI data to extract layer-specific connectvity patterns between resting state networks of the neocortex and using graph theory calculate spatial embeddings that provide differential connectivity patterns for each resting state network and indicate a global feedback state of the cortex at rest.