Professor Yuval Nir, a renowned sleep researcher at Tel Aviv University, is leading a groundbreaking study to optimize the early detection of dementia and improve outcomes for neurological conditions. Supported by a recent grant from Corundum Neuroscience, his project will investigate a machine learning-based approach to non-invasively detect abnormal brain activity in deep brain regions during sleep.
EEG has been a cornerstone of sleep research for nearly a century, but recent advancements in technology and data analysis are enabling new possibilities. We sat down with Professor Nir to discuss how AI and machine learning are unlocking deeper insights from EEG data, the unique benefits of sleep-state EEG for understanding brain health, and the potential for these advancements to accelerate the diagnosis and treatment of neurological disorders.
Electroencephalographic (EEG) measurements of sleep have been utilized for nearly a century. What recent advancements have been made in this field that change what how it is being used in research and elsewhere?
Prof. Nir: While EEG itself has indeed remained largely the same for nearly a century now, the most significant recent advance in my view is its powerful combination with AI and machine learning. There is much more information in EEG signals than what can be with visual inspection, and AI allows us to extract these signals with great sensitivity. In addition, in terms of hardware, “dry” electrodes (i.e. electrodes that measure the EEG without applying conductive gel) have improved, enabling better quality monitoring in people’ homes – an ambulatory setting – and not only in sleep laboratories.
While traditional research has focused on patterns of brain waves and their functions, your work explores a different type of electrophysiological events. Could you elaborate on paroxysmal discharges, their origins, and what they might signify? Furthermore, how do you interpret their relevance to a broad spectrum of neurological conditions, psychiatric disorders and/or mental health conditions?
Prof. Nir: Sleep EEG includes a number of “signature” brain waves such as slow waves or sleep spindles, which can be seen in all individuals. In contrast, paroxysmal discharges are abnormal brain activity patterns related to excessive, hypersynchronous firing of neurons (i.e. groups of neurons firing together in an abnormally coordinated way). They are most prevalent in epilepsy patients, where they can be associated with seizures, but also occur regularly between seizures as “interictal” activity.
These unique sharp waves are particularly noticeable during sleep and are observed not only in individuals with epilepsy, but also several other neurological disorders such as in neurodegeneration and dementia, in autism and ADHD, and after traumatic brain injury. They often originate from deep brain regions such as the hippocampus and medial temporal lobe, and are therefore difficult to detect with non-invasive EEG. Their relevance lies in their association with cognitive impairment, such as deficits in memory and language, and often suggest a poorer prognosis, for example more rapid cognitive decline in Alzheimer’s disease.
Nearly any neuropsychiatric disorder is associated with abnormal sleep – be it psychiatric disorders such as depression and anxiety or neurological disorders such as epilepsy, Alzheimer’s disease or Parkinson’s disease.
In what ways does sleep-state EEG provide unique insights that cannot be obtained from wake-state EEG? Specifically, how might it elucidate prodromal (early) states of various neurological conditions?
Prof Nir: Sleep serves as a powerful window into how typical and healthy brain activity of an individual is at a given time, both with respect to the general healthy population and also relative to the individual’s usual brain activity. One reason sleep is a good way to gain insight is that when we sleep, our brain goes through stereotypical stages, each with its own signature waves and patterns, making it easier deviations from these patterns. Indeed, nearly any neuropsychiatric disorder is associated with abnormal sleep – be it psychiatric disorders such as depression and anxiety or neurological disorders such as epilepsy, Alzheimer’s disease or Parkinson’s disease.
In addition, beyond sleep being a unique scientific and medical opportunity for monitoring brain health, it also offers technical advantages for better EEG monitoring, as it provides uninterrupted opportunity to record EEG for long hours without eye movements, blinking, and movements that interfere with the clarity of the EEG signal when we are awake.
How do you envision the future integration of sleep technology into daily life? What potential applications and implications arise from the development and widespread availability of accurate, non-invasive sleep monitoring devices for consumer use?
Prof Nir: Ultimately, I imagine that sleep will be monitored routinely at peoples’ homes with touchless monitoring devices that track physiological and brain activities. Combined with AI, it would be possible to construct a model for the typical sleep profile of each individual, and a significant deviation from this pattern would prompt people to consult with their physician for more detailed medical examinations.