Research

Transdiagnostic imaging-based markers of psychopathology

Most neuropsychiatric disorders are of neurodevelopmental origin, and their study and conceptualization continue to present with significant challenges, among them the lack of biologically-based indicators for distinguishing often co-occurring psychiatric disorders. In my line of work, I used a multimodal transdiagnostic approach that combines cutting-edge neuroimaging, clinical, and behavioral assessments. More specifically, I seek to extract latent dimensions that relate psychopathology to structural and functional brain alterations using magnetic resonance imaging and data-driven machine learning tools.

Relevant publications:

Development of imaging-based disease biomarkers

Neuroimaging data is becoming increasingly high dimensional and multi-faceted, suggesting that technique may contain clinically useful information for diagnostics, prognostics, and treatment monitoring. In previous studies, I utilized supervised machine learning (e.g., support vector machine and random forest classifier, kernel ridge regression) on structural and functional MRI data to predict to predict cognition, personality and mental health (Chen*, Tam*, et al., 2022), autism spectrum disorder (Benkarim et al., 2021), functional neurological disorder (Kebets*, Wegrzyk*, et al., 2018), and individual diagnosis of Alzheimer’s disease (Kebets et al., 2015; Kebets et al., 2018).

Relevant publications:

Identify brain-behavior relationships

I have collaborated on several projects that leveraged partial least squares analysis - an unsupervised machine learning technique - to identify various brain-behavior relationships :

Relevant publications: