The world smells different in Parkinson’s

People with Parkinson’s do not just lose their sense of smell; they actually perceive and experience aromas in a completely distinct way compared to everyone else. While scientists have known for years that a faded sense of smell is an early sign of Parkinson's, standard medical tests usually measure how well someone can identify or detect a scent. The problem is that these traditional tests cannot easily tell the difference between Parkinson’s and everyday smell loss caused by ageing, lingering viral infections, or simple sinus issues. A team of researchers at the Weizmann Institute of Science set out to see if there was a clearer way to spot the difference. Led by Michal M. Andelman-Gur and Noam Sobel, the team looked beyond basic performance to map out what they call an olfactory perceptual fingerprint. This unique metric measures two things together: how a person subjectively describes different aromas and how their automatic breathing patterns react while taking a sniff. The study brought together three distinct groups to compare responses. It included 33 people with Parkinson’s, 33 healthy individuals of a similar age, and 28 people who had lost their sense of smell due to other, non-Parkinson's causes. When given standard clinical smell tests, both groups with impaired smell scored almost identically, making them impossible to separate on paper. But when the researchers examined their perceptual fingerprints, a striking phenomenon appeared in how participants naturally snuffed up scents. When healthy participants or those with ordinary smell loss encountered a foul or unpleasant odour, their brain instantly triggered a protective reflex. They automatically shortened their sniffs by roughly eleven to twelve per cent to limit exposure to the nasty stench. In contrast, people with Parkinson’s showed a paradoxical twist in behaviour: when presented with unpleasant smells, they actually lengthened their sniffs, inhaling the unpleasant scent for longer instead of pulling away. By feeding these precise perceptual fingerprints and sniffing patterns into machine learning models, the researchers were able to separate people with Parkinson’s from those with unrelated smell loss with up to 88 per cent accuracy. When the models strictly matched individuals by age and sex, that accuracy jumped to 94 per cent. These discoveries show that Parkinson’s alters the fundamental way the brain processes and reacts to olfactory signals. Capturing these subtle, unique shifts in perception could pave the way for far more accurate, non-invasive diagnostic tools in the future.