Some of the most rewarding work I do is helping students carry out their first real research project. I mentor high school, undergraduate, and early graduate students one-on-one in computational neuroscience, AI, and data science. Over a project, each student goes from a broad interest to a focused question, analyzes real data or literature, and writes a complete research paper.
Below are snapshots of projects my students have completed or are working on now. They span brain imaging, machine learning for medicine, genetics, and cognitive science.
Combined language-model labeling of patient-written narratives (WHO functioning categories and sentiment) with numeric symptom scales into a 0–100 Wellness Index. He validated the labels against human annotators and tracked simulated patients over six months, then went on to explore the approach further with real patients.
Published · Young Scientist Journal (Vanderbilt), 2026 Read the paper →
Built multimodal encoding models that predict fMRI responses to movies from video, audio, and language features (Algonauts 2025 Challenge). He then used each region's functional-connectivity strength to identify a compact core of brain regions, overlapping known multimodal hubs such as the superior temporal sulcus, that carries most of the predictive signal.
Poster, CCN 2025 (Amsterdam) · First-author bioRxiv preprint Read the preprint →
Built an open-source, modular fMRI pipeline controlled by a single configuration file, covering quality control, regional homogeneity, static and dynamic connectivity, ICA, and graph metrics. She validated it on public OpenNeuro data and tested how sensitive results are to preprocessing choices.
Completed paper · open-source code
Extracted longitudinal MRI radiomic features from public glioblastoma scans (TCIA). He tested which features stay stable when the tumor segmentation is eroded, dilated, or shifted, using intraclass and rank correlation across four MRI modalities.
Completed paper
In-silico simulations comparing disruptive (NHEJ) and stabilizing (HDR, CAG→CAA) edits on synthetic HTT alleles. HDR shortened the uninterrupted CAG tract without changing polyglutamine length, moving alleles toward lower-risk categories.
Completed paper
Designed and ran an online survey that paired an emotion-vocabulary inventory with face-rating tasks, to test whether a richer emotional vocabulary changes how people perceive the intensity and valence of others' emotions.
Completed paper
A review of randomized trials, meta-analyses, and guidelines comparing lifestyle and drug interventions. Multidomain lifestyle programs showed the strongest evidence, and the review weighs the promise and the access barriers of new anti-amyloid antibodies.
Completed paper
A structured review of behavioral and neuroanatomical studies on whether lifelong bilingualism builds cognitive reserve that delays dementia symptoms. It treats the evidence carefully, as association rather than proof of protection.
Completed paper
A PRISMA-guided systematic review and meta-analysis that pools reported diagnostic accuracy (AUC) to compare blood-based and CSF-based biomarker panels paired with machine-learning classifiers.
In progress
A review drawing together structural MRI, diffusion imaging, and resting-state and task fMRI. It asks whether executive-function deficits in ADHD reflect delayed maturation of brain networks rather than a single damaged pathway.
Completed paper
Simultaneous recordings from a smartphone camera app he developed and a Polar H10 chest strap. The study tests how accurately short phone recordings measure HRV (RMSSD, SDNN) across adult age groups.
Data collection complete · analysis in progressFigures are illustrative schematics of each project's design, not reported results.
I mentor students independently and through these research programs:
Tell me a little about your background and what you're curious about.