Detecting Dementia from Patient Speech

by Martyn Allen in Guildford, England, United Kingdom

Detecting Dementia from Patient Speech


£5,000 target 11 days left
6% 11 supporters
Flexible funding – this project will receive all pledges made by 3rd November 2020 at 6:32pm

My name is Martyn Allen and I am just about to embark on a PhD with the University of Surrey in detecting dementia through segments of a pat...

by Martyn Allen in Guildford, England, United Kingdom

My name is Martyn Allen and I am just about to embark on a PhD with the University of Surrey in detecting dementia through segments of a patient's speech. The aim of the project is to combine speech analysis with advanced deep learning techniques to detect neurodegeneration at the earliest possible stage of it's natural progression. Dementia is a debilitating, elusive and persistent neurodegenerative disease. One that not only highly impacts the daily activities of those affected but also eradicates emotional memories with no hope of retrieval. A common misconception surrounding the disease is that it only affects the elderly age demographic. This, however, is very far from the truth. In the United Kingdom alone, there are believed to be over 40,000 individuals between the age of 30 and 65 that have been diagnosed with the disease. As of writing, there is no cure for neurodegeneration, of any form. For these reasons, among many, one can clearly appreciate the extensive research surrounding the treatment and diagnosis of dementia. Medical professionals across the globe are still using an outdated and highly inaccurate method of diagnosing the disorder. Often proving exceptionally time consuming and financially unfeasible, there can be no contestation that this area is in dire need of a radical solution. Imagine the following; an individual with suspected dementia enters a medical consultation with a professional and a conversation starts regarding daily activities and health. After a few moments, the medical professional reaches into the drawer of their desk and produces a small, seemingly insignificant device. The patient is asked to enunciate a specific sentence as clearly as they can through the recording feature of the device. Within seconds, the instrument lights up, generating a classification with one hundred per cent accuracy on whether or not the patient suffers from neurodegeneration.

I am aware that this is a large amount of money to raise, but every single penny will go a very long way towards my research. It will help support my course fees over the four-year duration, plus, and more importantly, the necessary equipment to ensure its success. Any help at all really would be greatly appreciated! Thank you again everyone, even for your interest, please please share as much you can!!

Let's make 'Detecting Dementia from Patient Speech' happen

Payment and personal details are protected