Reinventing dementia detection with AI

Cognetivity Neurosciences (CSE:CGN) is a Cambridge University spinout that has developed an AI-enabled digital health platform for cognitive assessment and early diagnosis of dementia. Applying the philosophy of “you can’t improve what you can’t measure”, they set out to apply recent advances in AI and Machine-Learning to the context of cognitive assessment. They hope that such a digital platform shall allow more people to be more aware of their cognitive health, and adapt their lifestyle in such a way that reduces the risks of cognitive decline that might come with age.

Health care is all about making complex decisions, particularly when it comes to secondary healthcare, which goes above and beyond screening and referral. Humans are better at caring but AI can do the repetitive tasks better; it can be more accurate; it can learn over time using huge amounts of data and therefore greater experience

What Cognetivity are doing is gathering the data and training an AI engine using this expertise to make a more informed diagnosis. As of today, there is no standalone diagnosis tool for dementia, given the complex nature of the disorder. Cognetivity’s platform does not remove the clinician from the equation, however, rather complementing them by providing them with the insights powered by the AI engine. This allows for more accurate diagnosis and more personalised treatments, which allows humans (puny humans!) to take better care of the patients.

Whereas conventional non AI-integrated cognitive assessments are subject to various biases: linguistic, learning and educational, Cognetivity’s ICA (Integrated Clinical Assessment) takes steps to try and remove these biases. Linguistic bias occurs when the performance of a test-taker is affected by the language that they speak. Cognetivity’s ICA therefore only uses pictures and symbols, thus removing this bias from their test. Learning bias is possibly the biggest flaw of conventional tests in that people who take the test multiple times naturally become accustomed to its format and thus improve with time, masking any degenerative effects on cognitive performance. Cognetivity’s co-founder has himself conducted a study into his company’s ICA and the subsequently published paper’s results show that the ICA neither suffers from learning bias nor educational bias (where the educational level of the test-taker affects their performance).

What about the future of AI in healthcare? Cognetivity’s co-founder predicts an increasing emergence of complementary AI, particularly in secondary healthcare. With the advent of new wellness technologies using AI, he also predicts a greater emphasis on prevention and lifestyle changes through the wealth of data and insights available to current users, calling upon the ability of even present-day fitness apps in providing a whole host of individual data points. The COVID-19 pandemic has actually accelerated this shift towards remote assessment and diagnosis. Nobody knows what the future holds for AI and healthcare but the use of AI in healthcare as an early diagnostic tool shall only grow, with Cognetivity at the forefront of the pursuit of faster, more accurate and more cost effective means of detecting degenerative diseases such as Dementia.

A huge thank you to Seyed from Cognetivity for sitting down and being interviewed by us. This article has not been sponsored in any way.