My primary clinical research interest is therapy for treatment-resistant neuropsychiatric disorders, with a focus on developing advanced brain stimulation and other neurosurgical techniques. I am among the pioneers of deep-brain stimulation (DBS) research in psychiatry.
1.I served as psychiatrist PI in a prospective, multi-center, double-blind, randomized and controlled European clinical study which investigated the impact of DBS treatment in depression. This study used DBS for treatment-resistant depression (n=5) for the first time in Israel. We found improvement in all patients treated with subcallosal cingulate DBS after six months with either high- or low-frequency stimulation (similar improvement was seen with both frequencies), and good long-term clinical results at the end of 12 months.
2. I was the first in Israel to use DBS in successful treatment of patients with obsessive compulsive disorder (OCD). Part of an international study, our results with 30 OCD patients, treated with internal capsule DBS, showed significant clinical improvement, and emphasized the importance of precise electrode location. Meantime, I have initiated an independent study of subthalamic nucleus (STN) DBS for patients with OCD and depression. We have also recently implanted two OCD patients with the new Percept brain pacemaker (Metronic LTD), enabling a first model of adaptive DBS stimulation.
3. I am PI in a study of ExAblate MR-guided focused ultrasound (FUS) bilateral anterior capsulotomy for OCD, developing with colleagues at Harvard and Stanford Universities the first evaluation of ExAblate MRgFUS for OCD, which is in the USA. We are currently using a two-stage research protocol to establish the safety and efficacy of ExAblate MRgFUS for patients with treatment-resistant OCD: stage 1 is designed as a ‘patient as own control’ paradigm (n=10) in two centers (Harvard and Stanford); stage 2 is a double-blind, randomized, control trial paradigm (n=56) in multiple centers.
With senior collaborator, neurobiologist Prof Hagai Bergman, we have developed novel methods and models for human deep-brain functional electrophysiological experiments. In parallel with our clinical studies, we are investigating the spectral signature of the emotional-cognitive basal ganglia in normal non-human primates engaged in emotional behavioral tasks, and in non-human primate models of emotional dysregulation and schizophrenia.
1. We have shown a right-left lateralization of emotional functions in the basal ganglia in Parkinson's patients. Our study used microelectrodes to record STN spontaneous spiking activity and responses to vocal non-verbal emotional stimuli during DBS surgery in these patients. We found the ventral regions to be asymmetrically associated with emotional function, and the right ventral STN linked with emotional processing.
2. We have studied cognitive aspects of the mechanism for movement facilitation and inhibition in the STN in parkinsonian patients performing increasingly complex oddball paradigms - auditory, simple movement and movement inhibition tasks. We found that the human STN responds principally to tasks involving movement - movement execution at the motor (dorsal) STN and movement planning at the limbic-associative (ventral) STN. We suggest that planning of every movement is routinely processed in the limbic-associative domain and is not limited to the motor domain. This implies that the STN sub-domains overlap and share functions, rather than work independently. Neuromodulation of these limbic-associative domains of the STN has generally been avoided, but it may, in fact, impact movement in a meaningful way.
3. Using a novel long-term deep electrophysiological recording device, we found a new sub-cortical circuit-level mechanism that generates obsessions and compulsions. We studied long-term STN electrophysiology in human OCD patients undergoing DBS and compared them with matching controls of Parkinson’s disease patients. Our investigational device, Activa PC+S (Medtronic Inc.), not only has standard stimulation capability, but also records long-term local field potentials. This study is the first to record long-term deep brain electrophysiological activity in psychiatric patients. It showed that theta (6.5-8Hz) oscillatory activity in the limbic-associative (ventral) STN inversely correlates with symptom severity and is modulated by provocation of OCD symptoms and other emotional and cognitive tasks. Surprisingly, we found similar ventral STN theta-alpha oscillatory activity in some 100 Parkinson’s patients who underwent DBS surgery. We thus suggest that theta-alpha oscillations can serve as an electrophysiological marker for the ventral STN limbic subarea and can guide optimal electrode placement in neuropsychiatric STN-DBS procedures, providing reliable biomarker input for future closed-loop DBS devices.
4. In this same study, we found beta (25-35Hz) oscillatory activity in the motor (dorsal) STN in OCD patients. This finding challenges the concept that these beta oscillations are pathognomonic for movement disorders and has triggered a search for a new concept for the closed-loop Parkinson’s DBS paradigm.
We then recorded neural activity in the cortical-basal ganglia network of normal non-human primates (NHP) while acutely and chronically modulating dopamine levels, up and down. We also assessed changes in beta oscillations in Parkinson’s patients following acute and chronic changes in dopamine tone. Beta oscillation frequency was strongly coupled with dopamine tone in both NHPs and human patients. In contrast, dopamine levels did not systematically regulate power, coherence between single-units and LFP, spike-LFP phase-locking or phase amplitude coupling (PAC). These findings demonstrate that it is frequency, rather than any other property, which is the key to pathological oscillation in the cortical-basal ganglia networks.
5. With PhD student Daniel Sand, we have developed a machine-learning (ML) algorithm to identify STN biomarkers which may enable automatic closed-loop STN stimulation. We have shown that individual patients possess unique sets of STN neurophysiological biomarkers, which can be detected over long periods. ML models have revealed that personally classified engineered features most accurately predict OFF vs. ON levodopa states. We have also recorded postoperative EEG cortical activity, while STN low frequency stimulation (LFS) was applied to different areas in- and outside the STN. ML models were used to differentiate the stimulation locations, based on EEG analysis of engineered features. A two-class linear support vector machine (SVM) predicted the interior (dorso-lateral region (DLR) and ventro-medial region (VMR)) vs. the exterior (Zona Incerta (ZI)) STN stimulation classification. Its accuracy was 0.98 and 0.82 for ZI vs. VMR and ZI vs. DLR, respectively, and 0.77 for DLR vs. VMR. Multiclass linear SVM predicted all areas with an accuracy of 0.82 for the exterior and interior STN stimulation locations (ZI vs. DLR vs. VMR). We suggest that noninvasive EEG biomarkers can use low-frequency STN stimulation to localize STN DBS electrodes to STN sub-regions. Such models can be used for both intraoperative electrode localization and postoperative stimulation programming, with potential to improve STN DBS clinical outcomes. Future development of adaptive DBS for Parkinson’s patients may include personalized ML algorithms.
6. We are studying different primate models for depression and schizophrenia. With MD-PhD student Maya Slovik, we have recorded local field potentials simultaneously from the primary motor cortex and the external globus pallidus of four monkeys, before and after administration of ketamine (NMDA antagonist). This study is the first to show spontaneous gamma oscillations under NMDA antagonist in non-human primates. These oscillations appeared in synchrony in the cortex and the basal ganglia. Phase analysis refuted the confounding effects of volume conduction and supported the funneling and amplifying architecture of the cortico-basal ganglia loops. These results suggest an abnormal network phenomenon with a unique spectral signature which could account for pathological mental and neurological states. With MD-PhD student Shiran Katabi, we have found neurophysiological and anatomical changes in the basal ganglia after treatment with antidepressant (Prozac) and destruction of the serotonergic neurons in monkeys. Our collaborative study with Prof. Suzanne Haber (Rochester University and Harvard University) has been awarded a BSF grant.
