Skip to NavigationSkip to content

AI and neurology: How machine learning is revolutionising neuroscience

Published on 28/09/21 at 04:05pm
Kat Jenkins

 

Artificial intelligence (AI) has undoubtedly been a growing presence in the healthcare industry, shaving years and billions of pounds off drug development programmes, accurately predicting A&E influxes, and even detecting early signs of disease in patients years before it was thought possible. The field of neuroscience has been no exception to this wave of technological innovation, with exciting developments cropping up in recent months and years that could potentially revolutionise diagnoses, treatments, and outcomes for patients on a global scale.

AI and subfields explained

 

The term AI covers a field of computer science that is focused upon the simulation of human intelligence and computational processes. However, there are several subfields of AI technology currently being explored in neuroscience, including machine learning (ML) and deep learning (DL). AI covers all programming systems that can perform tasks which usually require human intelligence. ML and DL have the same capabilities, with ML being a subfield of AI, and DL a subfield of ML.

There are many techniques and approaches to ML, one such approach is artificial neural networks (ANN). ANN are models of human neural networks that are designed to help computers learn, but more importantly are helping neuroscientists understand how the human brain functions in more detail than ever before.

AI and neurodegeneration

Alzheimer’s is a cruel and unpredictable disease, a type of dementia that corrodes away at the memory, thoughts, and behaviours of its victims. Diagnosis of Alzheimer’s is fiendishly complicated and often only achieved once the disease has already taken hold, which makes treatment difficult. In the UK it can take several scans and tests to diagnose, and often is difficult to distinguish between other forms of dementia. However, this could be set to change. A research team, led by Professor Zoe Kourtzi, a Fellow at the Alan Turing Institute and a Professor of Computational Cognitive Neuroscience at the University of Cambridge, has developed ML tools that can detect Alzheimer’s and dementia in patients at a very early stage. Their ML algorithm was trained using brain scans from patients who went on to develop Alzheimer’s, meaning it was able to learn to detect structural changes in the brain. When combined with the results from standard memory tests, the algorithm was able to provide a prognostic score – that is, the likelihood of the individual having Alzheimer’s disease. For those patients presenting with mild cognitive impairment – signs of memory loss or problems with language or visual/spatial perception – the algorithm was over 80% accurate in predicting those individuals who went on to develop Alzheimer’s disease. Amazingly, it was also able to predict how fast their cognition will decline over time.

Kourtzi explained, “We have trained machine learning algorithms to spot early signs of dementia by looking for patterns of grey matter loss – essentially, wearing away – in the brain. When we combine this with standard memory tests, we can predict whether an individual will show slower or faster decline in their cognition. We’ve even been able to identify some patients who were not yet showing any symptoms, but went on to develop Alzheimer’s. In time, we hope to be able to identify patients as early as five to ten years before they show symptoms as part of a health check”.

Dr Timothy Rittman from the Department of Clinical Neurosciences and a consultant at Addenbrooke’s Hospital, part of Cambridge University Hospitals (CUH) NHS Foundation Trust, is now leading a trial to look at whether this approach is useful in a clinical setting. So far, around 80 patients have taken part in the trial, which was run by CUH, Cambridgeshire and Peterborough NHS Foundation Trust, and two NHS trusts in Brighton. There are currently very few drugs available to help treat dementia. It is also thought that once a patient has developed symptoms, it may be too late to make a major difference. Thus, having the ability to identify individuals at a very early stage could help researchers develop new and effective treatments. If this trial is successful, the algorithm could be rolled out to thousands more patients across the country.

Aside from MRI data scanning, AI has also been used to great effect in testing frameworks in the field of neurodegeneration. For example, researches from Harvard Medical School (HMS) and Massachusetts General Hospital have this year developed an AI-based method to screen currently available medications as possible treatments for Alzheimer’s disease. The team created a ML algorithm framework called Drug Repurposing In AD (DRIAD) that works by measuring what happens to human brain neural cells when treated with a particular drug. The method then determines whether the changes induced by the drug correlate with molecular markers of disease severity. The approach also allowed the researchers to identify drugs that had protective as well as damaging effects on brain cells. One of the researchers behind the project, Artem Sokolov, explains, “Repurposing FDA-approved drugs for Alzheimer’s disease is an attractive idea that can help accelerate the arrival of effective treatment, but unfortunately, even for previously approved drugs, clinical trials require substantial resources, making it impossible to evaluate every drug in patients with Alzheimer’s disease. We therefore built a framework for prioritising drugs, helping clinical studies to focus on the most promising ones.” DRIAD also allows researchers to examine which proteins are targeted by the most promising drugs and whether there are common trends among the targets, an approach designed by Clemens Hug, an HMS associate in therapeutic science in the Laboratory of Systems Pharmacology and a co-first author of the paper that was published in Nature Communications in February. The team have now applied their screening method to 80 FDA approved clinically tested drugs, which has yielded a ranked list of potential candidates, with the top contenders including several anti-inflammatory drugs used to treat rheumatoid arthritis and blood cancers. These drugs belong to a class of medications known as Janus kinase inhibitors and work by blocking the action of inflammation-fuelling Janus kinase proteins, suspected to play a role in Alzheimer’s disease and known for their role in autoimmune conditions. The team’s analyses also pointed to other potential treatment targets for further investigation. One such drug, baricitinib, is now set to be investigated in a clinical trial for patients with subjective cognitive complaints, mild cognitive impairment, and Alzheimer’s.

AI and paralysis

One of the most astounding developments in neuroscience with the use of AI, is in the realm of paralysis. Several research teams are currently working with AI to help create mind-controlled prosthetics and even restore movement in paralysed humans. Chethan Pandarinath, a biomedical engineer at Emory University and the Georgia Institute of Technology, both in Atlanta, has been working to identify the patterns of electrical activity in neurons that correspond to a person’s attempts to move their arm in a particular way, so that the instruction can then be fed to a prosthesis. Essentially, he wants to read people’s minds. He fed his brain-activity recordings to an ANN, and tasked it with learning how to reproduce the data. The effort revealed the brain’s temporal dynamics – the way that its pattern of neural activity changes from one moment to the next – thereby providing a more fine-grained set of instructions for arm movement. He explains, “Now, we can very precisely say, on an almost millisecond-by-millisecond basis, right now the animal is trying to move at this precise angle... that’s exactly what we need to know to control a robotic arm.” It is important to note that there are several ethical concerns and restrictions to this kind of research, as there are questions over how much scientists can intervene in processes of the healthy human brain. Many recordings of neural activity in people come from the brains of those with epilepsy who are due to have brain tissue removed. This is because it is permissible to implant electrodes in brain tissue that will be excised anyway. Animal models do enable researchers to use more invasive procedures, but there are human behaviours, notably speech, that cannot be replicated in other species. Therefore, AI systems that can mimic human brain behaviours and not cause ethical concerns, such as ANNs, could be groundbreaking for neuroscientists.

Researchers from Battelle Memorial Institute and Ohio State University have taken this one step further and created a minute chip that, once implanted into the brain, can pick up electrical signals and transmit them to an arm cuff that contains specialised electrodes that stimulate the muscles and cause specific movements. In essence, they have created technology, through the use of AI, that enables brain signals to bypass damaged areas of the nervous system to allow the brain to communicate directly with muscles. The system is called the Battelle NeuroLife Neural Bypass Technology, and was created using ML brain-computer interface (BCI) neurotechnology back in 2016. In their study, the team implanted the chip in the motor cortex area of the brain of a partially paralysed 24-year-old man, and were able to decode the neuronal activity and control activation of his forearm muscles. Initially, he was able to pick up and hold a spoon using his own thoughts. After two years into the clinical trial, his level of function progressed to picking up and transferring objects, stirring liquids, swiping a credit card, and even playing a guitar video game. In May 2020, the team at Battelle reported that through the same technology, a sensorimotor neural interface successfully restored touch sensation in a patient with quadriplegia resulting from spinal cord injury. The NeuroLife team at Battelle are now working on an at-home BCI system, and have already been able to create a tablet-controlled computer system that can control muscle stimulation to elicit hand movements during home activities.

The depth of work currently underway in neuroscience thanks to AI is astounding. From the rapid detection and treatment curation for Alzheimer’s, to creating technology that can bypass damaged nerves and allow a patient to move paralysed limbs with their mind alone. It seems there are no limits to what this form of technology can achieve. However, when it comes to AI in neuroscience it is not just about curing ailments and diseases, it is also about coming closer than ever before to gaining a true understanding of the human brain itself.

Mission Statement
Pharmafile.com is a leading portal for the pharmaceutical industry, providing industry professionals with pharma news, pharma events, pharma service company listings and pharma jobs,
Site content is produced by our editorial team exclusively for Pharmafile.com and our industry newspaper Pharmafocus. Service company profiles and listings are taken from our pharmaceutical industry directory, Pharmafile, and presented in a unique Find and Compare format to ensure the most relevant matches