AI And The Future Of Mental Health

Doctors have moved one step closer to curing life-threatening neural diseases


There has been a lot of research and experimentation done to try and understand incurable brain diseases. Despite finding ways of easing the symptoms, prevention, and early diagnosis remain a problem that holds back a potential cure. Early detection of cognitive changes can significantly slow down disease development, its complications, and prevent devastating consequences. Based on cognitive science and being capable of processing huge amounts of data, artificial intelligence can be that missing element in solving a mystery around neural disorders.

Despite the progress in medical technology, when it comes to diagnostic tools for cognitive disorders, most of the time, healthcare specialists apply tests using good old pen and paper. Montreal Cognitive Assessment (MoCA) and the Clock Drawing Test (CDT) remain among the most effective to detect degenerative changes in the brain. With CDT, for example, a person is told to draw a clock face showing a specified time and then copy the previous drawing. The test demonstrates how people perform in terms of verbal understanding, spatial knowledge, and memory.

By using CDT, it is possible to say which potential disorder is currently affecting a person's brain. However, despite being of high accuracy, the method can still be subject to a doctor's subjective judgement when identifying to what extent cognitive changes have taken place. Thus, it's possible to say what type of disorder a person is suffering from, but unfortunately, only after the disease has started affecting the quality of life. In cases of Alzheimer's disease, for example, it can be decades before a cognitive change becomes noticeable.

While analyzing the issue, the Computer Science and Artificial Intelligence Laboratory at MIT (CSAIL) suggested that instead of analyzing final drawings or scores in cognitive tests like CDT and MoCA, scientists and doctors should consider the process of the assessment as a whole. In order to extract data from the process of assessment, the research team combined hardware and artificial intelligence and created the Anoto Live Pen. The digital ballpoint pen is capable of measuring its position on the paper upwards of 80 times a second, with a built-in camera capturing movements and collecting important data. Data collected from 2,600 tests over 9 years allowed the CSAIL team to create specialized software in the form of the Digital Clock Drawing Test (dCDT).

The newly developed method proved to be more accurate than the conventional CDT, thanks to the more thorough analysis of the results due to the thousands of new data points available. The CSAIL team has found that individuals with memory impairments spend more time thinking prior to drawing than those without a disorder. Also, individuals with signs of Parkinson's disease spend more time drawing the clocks - these insights were impossible to see with the conventional test. The main purpose of the dCDT is to save time on the detection of cognitive changes, so the disease can be diagnosed at earlier stages.

However, the CSAIL team is not the only one who believes that AI and its capabilities of learning from data can contribute to the diagnostics and treatment of degenerative diseases. A former founder of Braintree Bryan Johnson has recently invested $100 million in Kernel - a human intelligence company which is developing the world’s first neuroprosthesis to mimic and improve cognition of mentally ill patients.

The idea of the project is to create a prosthesis which can be either implanted or attached to the body, where machine learning algorithms facilitate communication between brain cells by operating directly in the neural code that is responsible for storing and recalling information in the human brain. Whilst the device’s priority is to understand and slow down the cognitive changes, researchers from Kernel also believe that the overall development of AI can benefit too.

Artificial intelligence is based on the deep understanding of how the human brain works, so if technology is in direct communication with the organ, the research and development of AI can be more accurate. However, due to ethical concerns around AI, it is likely that the interaction with the human brain and a machine will be difficult to implement any time soon.

AI is a field with enormous potential and despite numerous fears that one day, technology will act against humanity, so far, solutions provided by machine learning mostly act to the society’s benefit. AI driven approaches are already disrupting conventional methods of diagnostics, showing outstanding results like in case with dCDT. An increasing interest among investors and scientists to integrate AI within healthcare system also means that one day those who suffer may be able to regain their independence of mind.

Vision small

Read next:

Big Data Forecasting In Pharma