Abhay Koushik

Engineer. Neuroscientist

I research & develop systems that interpret brain, body & behaviour to support human well-being.

Working Territory

Complex signals. Human questions.

I work with EEG, neuroimaging, physiological rhythms, behavioural data, and full-body movement to investigate sleep, cognition, perception, and bodily regulation—and to build tools that make these processes more interpretable.

Abhay Koushik presenting a body–brain study design to an audience in 2024 — Brain–Body Waves ConferenceResearch Presentation · 2024

Across Research and Product

From exploratory code to working systems.

My work spans research laboratories, startup R&D, behavioural experiments, neuroimaging studies, and rapid prototypes. The recurring task is turning an uncertain question into an analysis, experiment, or system that can be tested.

Wearable EEG signals displayed across a laptop and mobile sleep-staging applications — AmuseMeWearable EEG and mobile sleep-staging prototype

Selected Work

From signal to system.

Research across wearable sensing, multimodal brain data, motion analysis, and human perception—built into working systems.

Wearable EEG sleep staging interface with hypnogram and signal visualization

Wearable EEG · On-Device Inference · Physiological Computing

Deep Learning–Based Real-Time Sleep Staging

Built an end-to-end smartphone system for real-time sleep staging from single-channel wearable EEG, combining EEG/IMU streaming, TensorFlow Lite inference, and power-spectral analysis.

ProblemCan sleep stages be inferred outside a full sleep-lab setup?

Outcome

Co-authored papers at IEEE BSN and the ML4H/NeurIPS workshop.
Selected for an oral presentation at IEEE International Conference of BSN.

Explore Project

Functional brain connectivity matrices filtered using PET-derived neurotransmitter receptor maps

Multimodal Neuroimaging · Pharmacology · Biomarker Development

Neurotransmitter-Informed Connectivity

Implemented a MATLAB pipeline connecting EEG and fMRI connectomes with normative PET maps, using quantile filtering, randomized spatial controls, and mixed-effects modelling.

ProblemCan neurotransmitter geography help explain functional connectivity and pharmacologically induced network change?

Outcome

Enabled completion of the multimodal pipeline implementation investigating drug effects as part of my lab rotation
Contributed as a co-author to the broader Neurobiological Organization of Functional Connectivity study.

Explore Project

Full-body HoloSuit motion capture for yoga training alongside reconstructed emotional faces for social cognition in virtual reality

XR Product R&D · Motion Intelligence · Human Sensing

Embodied Intelligence in XR

Built full-body XR modules for motion capture, skeletal calibration, gesture control, and trainer–learner movement comparison using quaternion kinematics and Dynamic Time Warping. Also reconstructed and validated 3D emotional faces for social-cognition experiments in virtual reality.

ProblemHow can an interactive system represent, personalize, and compare human movement and social cognition in real time?

Outcome

Added a Yoga motion-intelligence training module for HoloSuit's real-time training platform.
Developed 3D emotional stimuli for Social Cognition and VR research at ENS Paris.

Explore Project

MRI and PET illustration showing partial volume effects in the midbrain

Clinical Neuroimaging · PET/MRI · Neuroendocrinology

PMDD Serotonin PET Study

Produced MRI morphometry, ROI analyses, and PET × MRI visualizations during peer-review revision of a study on cycle-linked serotonin-transporter dynamics in Premenstrual Dysphoric Disorder (PMDD).

ProblemCan cycle-linked serotonin transporter changes help explain PMDD symptoms?

Outcome

Presented my poster the PMDD PET study contributions helping a positive evaluation of the Max Planck School of Cognition evaluation committee.
Contributed as one of the co-authors to the PMDD study published in Biological Psychiatry.

Explore Project

3D face stimuli and behavioral similarity matrix visualization

Social Cognition · Online Experimentation · Behavioural Modelling

FaceSim3D Social Cognition

Co-designed and implemented a Unity/UXF online experiment comparing face-similarity judgments across 2D and dynamic 3D stimuli, with representational similarity analysis.

ProblemDo people judge facial similarity differently when faces are presented as flat images or dynamic 3D stimuli?

Outcome

Co-authored a Royal Society Open Science paper on Face Perception.
Contributed to research posters presented at Mind Brain Body Sympsium and Computational Cognitive Neurosciences (CCN) Conference.

Explore Project

Award-winning prototypes spanning silent-speech sensing, generative AI, and automated microscopy cell counting

Rapid Prototyping · Computational Biology · Generative AI

Award-Winning Rapid Prototypes

Contributed to three rapid team builds spanning magnetometer-based silent-speech sensing, GPT-3 conservation storytelling, and OpenCV microscopy cell counting.

ProblemHow can emerging computational technologies be turned rapidly into useful, human-centred systems across health, biology, and conservation?

Outcome

Contributed to teams awarded at HackHarvard and Stanford TreeHacks.
Contributed to the Paris Bettencourt iGEM team awarded Gold and Best Manufacturing Project.

Explore Project

Global Community Bio Summit fellowship and wearable EEG audio feedback visual

Neurotechnology · Collective Intelligence · Global Leadership

Global Community Bio Leadership

Built an Android system translating wearable-EEG alpha activity into real-time audio, presented it as a Global BioSummit workshop, and later contributed to MIT–Takeda Superminds.

ProblemHow can physiological feedback, community biotechnology, and collective intelligence make health and self-regulation more tangible and accessible?

Outcome

Joined the organizing team of GCBS, later got selected as a Global Community Bio Fellow.
Contributed to MIT–Takeda Superminds work on collective intelligence for mental health.

Explore Project

Industry Direction

Rigorous research. Usable technology.

I work best where complex human data, scientific uncertainty, and technical development meet—turning difficult questions into evidence, prototypes, and systems that can be tested in the real world.

Award-winning hardware and generative AI prototypes developed at HackHarvard and Stanford TreeHacks — Award-Winning Rapid Prototyping
Best Hardware Hack · HackHarvardBest OpenAI API Hack · Stanford TreeHacks

Conversation

For research and technology that demand more than one way of thinking.

Start a Conversation View Trajectory