EXPLORATORY RESEARCH

We develop novel tools for assessment of mental wellbeing and cognition and analysis of the EEG signal, and use machine learning and AI frameworks to understand diverse human physiological, mental and cognitive phenotypes and their social and environmental determinants.

Brain Physiology

EEG is a non invasive, portable and cost effective neuroimaging technology that allows easy acquisition of large scale data and deep insights into brain physiology. Its primary advantage is its high temporal resolution. We develop novel tools to extract features of its rich temporal structure to better understand how the brain signal relates to cognitive and mental health outcomes.

Mental Health

Our work in mental health assessment has led to the development of a tool called the Mental Health Quotient or MHQ for the comprehensive assessment of mental wellbeing along a spectrum from clinical to thriving that is used in the Mental Health Million Project. Our research focuses on using this data to understand clinical and normal mental phenotypes in the population, their corresponding physiological or EEG profiles as well as social determinants of mental health outcomes.

Cognitive Health

There are many aspects to cognition. We work on ways to measure and assess its elements across the life span in a way that that can extend across languages and cultures and be readily related to brain activity.
Our research focuses on understanding cognitive phenotypes, EEG predictors of cognitive health across the lifespan and social and environmental drivers of cognitive expansion.

Featured research

Characterizing Peaks In The EEG Power Spectrum
Parameshwaran and Thiagarajan, Biomedical Physics & Engineering Express, June, 2019
Novel method to separately analyze the strength and fidelity of peaks in the power spectrum independent of the background decay that does not depend on the shape of the decay and curve fitting. Differences in Alpha Oscillations between eyes open and eyes closed assessed with this method are more consistent than other methods.
Waveform Complexity: A New Metric For EEG Analysis
Parameshwaran et al, J Neurosci Methods, Oct 2019
Waveform Complexity is better correlated with performance on a
pattern complexity task than related measures of spectral
entropy, sample entropy and lempl-ziv complexity.
The Heterogeneity of Mental Health Assessment
Frontiers in Psychiatry, 2020
This analysis of 126 commonly used mental health assessment tools spanning 10 psychiatric disorders reveals substantial symptom inconsistency within disorders as well as considerable overlap across disorders. These results highlight challenges with the current methods of assessment in enabling the understanding of underlying etiologies and the discovery of new treatments.
EEG Frequency Bands in Psychiatric Disorders: A Review of Resting State Studies
Frontiers in Human Neuroscience, 2019
This analysis of 184 clinical EEG studies, spanning 10 different mental health disorders reveals the lack of specificity of frequency band profiles (e.g. alpha, beta) across disorders, the considerable variability of results within disorders, and highlights the poor standardization in methods across studies. These results call out the need for standards to enable reproducibility of results, and novel EEG metrics that are better able to distinguish between disorders and symptom profiles.
Figure shows consistency and validity scores for the patterns of frequency bands across studies for each disorder. Longer bars = more consistent.
Development, feasibility and acceptability of a gamified cognitive DEvelopmental assessment on an E-Platform (DEEP) in rural Indian pre-schoolers – a pilot study
Bhavnani et al, Global Health Action, Jan 2019
A game based cognitive assessment for preschoolers tests a wide range of cognitive elements and was effective in rural preschoolers in India.