Machine intelligence benefiting treatment outcomes
At Rashid Hospital, the neurology team is exploring the different ways AI can benefit traditional patient care and support treatment outcomes. We are working with emerging technologies that utilise neuroscientiﬁc ﬁndings to inform machine intelligence-driven advances in the field of medicine.
For example, we are currently working with RAPID stroke imaging software, which helps radiologists and neurologists evaluate acute strokes using a sequence of three brain scans (CT, CT perfusion and CT angiography).
A radiologist usually needs around 40 minutes to analyse individual scan data, but this software reduces that to just two minutes; dramatically accelerating the diagnostic process.
As a community hospital, we are a hub for research implementation and, together with Dubai Government and Dubai Future Accelerators, there are a number of neuroscience projects that we are involved with.
These include an early phase neuro headband developed by a Canadian research team from Healthcare Innovation in Neuro Technology (HiNT) for early detection in the case of acute stroke and transient ischemic attack (TIA). The technology sends a message to the treatment team to inform them that the patient is having an acute stroke for rapid response.
Mind mapping to address neurological disorders
Globally, around 44 million people are diagnosed with Alzheimer’s or related forms of dementia. As we continue to map brain function and capability, neuroscientific research is providing a solid base for AI innovation with under-development assessment tools that may help patients’ degenerative visuospatial dysfunction.
A team from Massachusetts Institute of Technology (MIT) has developed pre-emptive software that can determine whether high-risk prospective Alzheimer’s patients will experience cognitive decline up to two years in advance.
The background research used clinically significant cognitive test scores and other biometric data from Alzheimer's patients and healthy individuals as a basis to create an analytical model that effectively learns patterns to assist in predicting how patients will score on cognitive tests.
Autism, which currently affects one in every 160 children according to the World Health Organization, is another key research area where our expanding knowledge of brain function can be coupled with AI advancements to address rising numbers of spectrum disorder diagnoses.
A 2019 Google Glass and Stanford University Medical School clinical trial determined that autism spectrum children who wore the wearable technology at home showed significant improvements in social skills.
The ‘Superpower Glass’ programme helps children classify the emotion of the person they are interacting with using machine learning that enables the identification of eight emotions. The child is then ‘cued’ via a robotic audio clip and visual emoticon and supported by a family companion app managed by the relevant caregiver.
Researchers at MIT Media Lab are also working with robots that use deep learning to estimate the unique engagement and interest of autism spectrum disorder children; personalising data specific to each child. The goal is not to replace human therapists with robots, but to provide augmented information in order to personalise therapy sessions and deliver better outcomes.
Next-level understanding needed to advance AI
Neuroscience definitely isn’t in its infancy; rather we’re at the nursery stage of understanding brain function, but we are still constantly surprised by how different cells inform different functions and affect quality of life.
Take Parkinson’s disease, which is a progressive disorder caused by degeneration of nerve cells in the part of the brain that controls movement whereby these cells die and lose the ability to produce the chemical messenger dopamine. These dopamine cells represent just 0.1 per cent of the total neuron count yet they have a hugely debilitating effect on movement and overall function, affecting an estimated five million people worldwide.
IBM’s fingernail sensor prototype project, funded through the Michael J Fox Foundation, is using AI to help clinicians track, monitor and more accurately diagnose Parkinson’s disease. The tech giant has also shared published research on the use of AI and machine learning to better detect and comprehend changes in patient’s speech, which can also indicate markers of Parkinson’s disease progression.
This is one way in which AI is proving to be invaluable in providing key insight into these types of neurological disease. The ability to use AI to analyse immense amounts of data collected from longitudinal, clinical, behavioural and imaging assessments observed from patients in medical settings, as well as genomic and biological samples, could be game changing.
Understanding how brain tissue disfunction influences neuropsychological disorders such as bipolar, depression, schizophrenia and psychosis could also benefit from AI advancements used to augment technological applications like functional MRIs and nuclear medicine imaging.
For example, researchers from the Chinese Academy of Sciences have developed an AI algorithm that works in tandem with functional MRI applications to help identify the consciousness recovery likelihood of brain damaged patients.
And a new deep learning algorithm developed by scientists from Emory University (US) and Harvard University, uses a machine-learning method that can predict individuals who may be at risk of psychosis with 93 per cent accuracy simply from hidden clues in speech.
Dubai as a hub for AI research
Dubai is at the forefront of AI advancement in the healthcare field both as a regional hub and as home to world-class hospitals and health centres. Scientists from around the world are choosing the emirate to continue their research, trial AI technologies and network with the international community at events such as this month’s World Congress of Neurology at Dubai World Trade Centre.
The scope and ambition of the world’s AI innovators coupled with the medical profession’s ongoing research into the hidden depths of brain function is enabling some world-changing technologies – but we are in it for the long haul. Government programmes such as Dubai Future Accelerators are bringing in fresh talent and research. As a stroke specialist, at the moment I’m excited to be working with a Korean company, developing software that will be able to read patient MRI images and specify whether an ischemic stroke is cardioembolic or something else.
This has the potential to offer cost-saving benefits, cut diagnosis times and save lives.
And here we come full circle. If you don’t understand your machine – in this case, the brain – it is difficult to determine exactly how you can help it. But, if we continue with the same degree of neuroscientific advancement in the next 20-plus years as we’ve seen in the last two decades, there is plenty to be excited about.