HIT at the Annual Meeting of ISFN – the Israel Society for Neuroscience

HIT at the Annual Meeting of ISFN – the Israel Society for Neuroscience

At the Annual Meeting of ISFN (Eilat, January 2025), HIT was represented by three outstanding researchers: Dr. Erez Simony, Senior Lecturer and Head of the MSc program at the Faculty of Electrical & Electronic Engineering; Dr. Gaddi Blumrosen, Senior Lecturer at the Faculty of Sciences, Dept. of Data Science, and Dept. of Digital Medical Technologies; and Dr. Hadas Lewy, Head of Digital Health Ventures and Senior Lecturer at the Dept. of Digital Medical Technologies. Their work aroused a great deal of interest at the conference, creating fertile ground for future collaborations with colleagues from other research institutions. 

New Scientific Discovery: the Movie After-Effect

Dr. Erez Simony, Head of the MSc with a Thesis program at HIT's Faculty of Electrical & Electronic Engineering, and advisor at the Dept. of Brain Sciences at the Weizmann Institute, presented an interesting study in the session 'Learning & Memory: From Mice to Humans', at the Annual Meeting of ISFN. The study was conducted as part of the MSc thesis of HIT student Nir Yahav, in fruitful collaboration with Prof. Rafael Malach from the Dept. of Brain Sciences at the Weizmann Institute. 

Dr. Simony: "I hope that more students with scientific curiosity and motivation to break new boundaries will join our research track – MSc with a thesis at HIT's Faculty of Electrical & Electronic Engineering. Our program offers a direct track from the fourth year of BSc studies, as well as excellence and subsistence scholarships for outstanding students."
The study presented at the conference was based on data from fMRI (functional MRI) scans – a technology that measures neural activity in the human brain while the subject performs a task. The scans were obtained from the Human Connectome Project (HCP) of the National Institutes of Health (NIH) in the US. 

Master's student Niv Yahav analyzed fMRI data from 170 participants, whose brains were scanned while watching 14 video clips lasting between one to five minutes, extracted from different movies, with 20 seconds of rest between each clip.. The study's first goal was to classify neural activity in the last five seconds of each clip, in 300 different brain regions, , in order to identify which movie the participant had watched.. To this end, the researchers used a machine learning model called SVM (Support Vector Machine), applying the leave-one-out cross-validation methodology. This means that the model was trained on 169 participants, and its prediction capability was tested on the 170th subject (and then in another run on all 170). The study found a high rate of accuracy in the model's predictions, especially in the brain's visual and attention regions, as well as in the Default Mode Network (DMN), associated with memory, planning, and contextual processes. 

Dr. Simony: "A great surprise awaited us when we ran the same model on the rest intervals between clips: The model was able to identify which clip had just been viewed based on the last five seconds of the rest period. As far as we know this is the first time researchers have been able to identify a clip based on brain activity during the post-viewing interval." These findings led to the question: What happens in the brain during the 20-second post-viewing rest period, which enables the model to identify brain activity patterns across the subjects? How does it manage to associate brain activity during rest with the movie that the participant just watched?

Dr. Erez Simony

Dr. Simony: "Here we discovered a fascinating phenomenon in all brain regions that reacted to the clips: brain activity during rest was the exact reverse of activity toward the end of the clip. In other words, the neurons most highly activated at the end of the clip, were suppressed during rest, while those that displayed low activity at the end of the clip – especially in the DMN – showed rebound activation during rest. To describe this phenomenon, we introduced a new term: The Movie After-Effect. 

Prof. Malach: "The phenomenon we discovered is related to homeostasis – a basic control mechanism that the brain employs to maintain constant balance. Our research is apparently the first to demonstrate the homeostasis phenomenon using natural stimuli across all regions of the human brain. The phenomenon points to the fascinating fact that our brain is constantly in a dynamic state, continuously changing its properties in accordance with the ever-changing environmental conditions."

In future studies the researchers intend to investigate whether the mechanism they have discovered plays a role in short-term memory – with a focus on people with cognitive decline or Alzheimer's disease. 

New Paradigm in Brain Science: Behavior is the Key

Dr. Gaddi Blumrosen of HIT initiated and led the session 'Monitoring and diagnostics of neurological diseases and disorders in home environment settings' at the Annual Meeting of ISFN. Dr. Blumrosen explains that this was a unique session, representing a revolutionary paradigm in neuroscience: "Instead of trying to monitor brain signals – a complex process, sometimes involving the insertion of invasive electrodes, we apply advanced AI tools to explore the behavioral implications of these signals. This approach enables us to diagnose and treat pathologies of neurological diseases and optimize the patients' wellbeing."

Dr. Blumrosen presented his proposed paradigm: using AI and signal processing to monitor and quantify behavioral parameters and build decision-support systems for both clinicians and patients. Such systems can facilitate effective personalized diagnosis and treatment, and also monitor the treatment's effectiveness. Focusing mainly on neurological diseases and mental conditions and their diagnostic biomarkers, Dr. Blumrosen collaborates with Israel's largest medical centers, including Sheba, Ichilov and Beilinson. 

Dr. Gaddi Blumrosen

Dr. Blumrosen and his colleagues combine advanced AI methodologies with sensors that are readily available in every home – in computer games, smart watches or simple videocams, to produce continuous reliable data about patients' behavior in their natural environment. In this way, he explains, "we can, for example, monitor the condition of a patient sent home after rehabilitation, diagnose depression from the subject's activity level, or assess the condition of people with essential tremor, Parkinson's disease, or dementia."

In his talk, Dr. Blumrosen demonstrated the implementation of the novel approach through a unique study on ADHD - an MSc project led by HIT alumnus Anton Gellashvili, who will soon complete his MSc with distinction at HIT's Dept. of Computer Science. Dr. Hezi Reshef, a data science expert from the Hebrew University Business School, also contributed to the study.  The researchers developed a complex AI model enabling the diagnosis of ADHD based on a short real-life video of the subject. To date, 80 individuals have been diagnosed in this way, with a success rate of 80%. The method has been copyrighted and published in scientific publications. 

Researchers from other institutions, most of whom collaborate with HIT, also presented their work at the unique session: medical sensing through a watch (Prof. Joachim Behar, the Technion); evaluating motor sensing of Parkinson's patients from how they play the piano (Prof. Jason Friedman, Tel Aviv University); Initial diagnosis of Parkinson's disease from online interactions (Dr. Inbal Maidan, Ichilov Medical Center); Diagnosing mental conditions from remote interviews (Dr. Hila Gvirts, Ariel University); and the Living Lab at HIT (Dr. Hadas Lewi, HIT). 

The novel approach presented by DR. Blumrosen and others aroused a great deal of interest at the conference, paving the way for the paradigm's implementation in different areas.

Digital tools for personalized preventive medicine

Dr. Hadas Lewy of HIT presented her research at the session 'Monitoring and diagnostics of neurological diseases and disorders in home environment settings' at the Annual Meeting of ISFN. Dr. Lewy's work focuses on developing digital technology-based diagnostic and monitoring tools, especially for elderly people with cognitive decline. She explains that the advanced tools can help treatment providers, such as geriatricians and occupational therapists, providing them with an objective means for measuring their patients' condition. 

Dr. Lewy and her colleagues conduct their research in two state-of-the-art labs on the HIT campus: the Living Lab designed as a home environment, where they develop innovative models for monitoring patients in their homes, and conduct clinical trials jointly with hospitals; and the App Lab developing easily accessible apps that combine widely used cognitive tests with smart diagnostic systems. Projects are conducted in collaboration with Clalit Health Services and the geriatric medical centers Beit Rivka and Herzfeld.

Dr. Hadas Lewy

In her talk at the conference Dr. Lewy explained that digital tools are crucial for the advancement of personalized preventive medicine – a main goal of present-day medicine. She presented an example of using such tools for cognitive diagnosis: "Today, when patients are sent home after geriatric rehabilitation, there is a need to assess their independence at home: can they warm their food, call their family or the doctor, etc., or do they require more support and supervision, or additional treatment?" Dr. Lewy hopes that in the future digital sensors will be installed in the home of every person who is at risk due to age, illness, or any other cause, to enable early diagnosis and intervention leading to effective personalized treatment.

Dr. Lewy concludes: "Digital diagnostic and monitoring tools have enormous potential. They allow us to reach wider populations, identify disease, aging problems and cognitive decline at an earlier stage, and provide treatment to prevent further deterioration. In addition, the huge databases formed through widespread digital monitoring will advance medical science by facilitating a better understanding of diseases and various phenomena."