Our lab explores the molecular mechanisms underlying brain disease and develops MRI-based methods for early diagnosis and therapy optimization, with potential applications in cancer, stroke, and neurodegenerative disorders. We design and implement machine-learning-based strategies for quantitative molecular parameters imaging.

Research Overview
Alterations of in vivo molecular properties, such as intracellular pH and protein/metabolite concentrations, are known to underlie a variety of neurological diseases, including stroke, traumatic brain injury, epilepsy, Alzheimer’s disease, and cancer. Importantly, pathology-related effects at the molecular level are manifested long before any anatomical or structural changes occur and may also reflect the host response to therapeutics. Therefore, non-invasive imaging of molecular processes is a pivotal tool for achieving early diagnosis and could play a crucial role in the optimization of therapeutic approaches.

Our main goal is the development of clinically relevant molecular MRI methods for early diagnosis and characterization of neurological diseases.

Identifying molecular bio-markers for tissue characterization
We develop tools for disentangling the different signals coming from brain metabolites, proteins, and lipids, and evaluate their potential to serve as noninvasive image bio-markers for cell death, ischemia, and disease severity.

Developing AI-based methods for accelerated and optimized molecular MRI
We design and implement AI-based methods for early interventions along the imaging pipeline, enabling automatic MRI acquisition protocol discovery and quantitative molecular parameters reconstruction. This allows for a drastic reduction in scan time and compatibility with a variety of biological scenarios.

Noninvasive image monitoring of treatment response
While conventional imaging provides anatomical information that reflects the treatment resopnse to therapy at a relatively late stage (e.g., in the form of tumor shrinkage), molecular MRI has the potential to inform the physician about the treatment response (or the lack of it) much earlier. This is crucial for optimizing novel therapeutic modalities, and for providing precision medicine. We investigate the suitability of particular brain metabolites to serve as early indicators of treatment response for a variety of pathologies.

Quantifying the neurochemical changes occurring in health & disease
We investigate the use of deep neural-networks and magnetic resonance fingerprinting (MRF) for rapid extraction of quantitative maps from chemical exchange saturation transfer (CEST) and semisolid magnetization transfer (MT) signals. Such quantitative parameters may improve the reproduciblity and robustness of molecular MRI, and facilitate its clinical acceptance.

For more information:
Email:
orperlman@tauex.tau.ac.il
Website:
https://mri-ai.github.io/
Address:
Tel-Aviv University 6997801, Israel
Location:
Multidisciplinary Research Building, Room 410
Phone:
+972-3-6409418