SmellSense: Feasibility of Detection of Mild Cognitive Impairment and Alzheimer’s Disease using Olfactory Stimulus and Electrophysiology
Pehel Mehta and Amal Jude Ashwin F
25/09/2026
Alzheimer’s is a neurodegenerative disease that affects roughly 8.8 million Indians over the age of 60. Mild cognitive impairment is an early stage of AD and is crucial for early diagnosis and intervention of AD. Unfortunately, early MCI is asymptomatic and is not sensitive to self-reported questionnaires and scans that are widely used for AD diagnosis. In this study, we propose to use olfactory stimuli in an oddball paradigm paired with electroencephalogram (EEG) to classify healthy, MCI and AD subjects. We developed a simple machine learning model that was able to classify the 3 classes with a high accuracy of >94% (33/35 subjects) validated using leave-one-subject-out cross-validation. Additionally, the real-world feasibility of the system was piloted using 5 healthy adult subjects using in-house devices, and interpretable summary reports were generated using Python. Overall, these results support the feasibility of SmellSense as a portable and cost-effective tool for detecting MCI and AD using olfactory stimuli and EEG, though further validation on larger and clinically diagnosed populations is needed before any diagnostic claims can be made.