Deep Brain Stimulation for Treatment-Refractory Mood and Obsessive-Compulsive Disorders

Controlled ultrasonic interventions through the human skull

Transcranial focused ultrasound enables precise and non-invasive manipulations of deep brain circuits in humans, promising to provide safe and effective treatments of various neurological and mental health conditions. Ultrasound focused to deep brain targets can be used to modulate neural activity directly or localize the release of psychoactive drugs. However, these applications have been impeded by a key barrier-the human skull, which attenuates ultrasound strongly and unpredictably. To address this issue, we have developed an ultrasound-based approach that directly measures and compensates for the ultrasound attenuation by the skull. No additional skull imaging, simulations, assumptions, or free parameters are necessary; the method measures the attenuation directly by emitting a pulse of ultrasound from an array on one side of the head and measuring with an array on the opposite side. Here, we apply this emerging method to two primary future uses-neuromodulation and local drug release. Specifically, we show that the correction enables effective stimulation of peripheral nerves and effective release of propofol from nanoparticle carriers through an ex vivo human skull. Neither application was effective without the correction. Moreover, the effects show the expected dose-response relationship and targeting specificity. This article highlights the need for precise control of ultrasound intensity within the skull and provides a direct and practical approach for addressing this lingering barrier.

Remotely controlled drug release in deep brain regions of non-human primates

Many areas of science and medicine would benefit from selective release of drugs in specific regions. Nanoparticle drug carriers activated by focused ultrasound-remotely applied, depth-penetrating energy-may provide such selective interventions. Here, we developed stable, ultrasound-responsive nanoparticles that can be used to release drugs effectively and safely in non-human primates. The nanoparticles were used to release propofol in deep brain visual regions. The release reversibly modulated the subjects' visual choice behavior and was specific to the targeted region and to the released drug. Gadolinium-enhanced MR imaging suggested an intact blood-brain barrier. Blood draws showed normal clinical chemistry and hematology. In summary, this study provides a safe and effective approach to release drugs on demand in selected deep brain regions at levels sufficient to modulate behavior.

Remotely controlled drug release in deep brain regions of non-human primates

Many areas of science and medicine would benefit from selective release of drugs in specific regions of interest. Nanoparticle drug carriers activated by focused ultrasound-remotely applied, depth-penetrating energy-may provide such selective interventions. Here, we developed stable, ultrasound-responsive nanoparticles that can be used to release drugs effectively and safely in non-human primates. The nanoparticles were used to release propofol in deep brain visual regions. The release reversibly modulated the subjects' visual choice behavior and was specific to the targeted region and to the released drug. Gadolinium-enhanced MRI imaging suggested an intact blood-brain barrier. Blood draws showed normal clinical chemistry and hematology. In summary, this study provides a safe and effective approach to release drugs on demand in selected deep brain regions at levels sufficient to modulate behavior.

Noninvasive modulation of essential tremor with focused ultrasonic waves

Objective: Transcranial focused low-intensity ultrasound has the potential to noninvasively modulate confined regions deep inside the human brain, which could provide a new tool for causal interrogation of circuit function in humans. However, it has been unclear whether the approach is potent enough to modulate behavior.Approach: To test this, we applied low-intensity ultrasound to a deep brain thalamic target, the ventral intermediate nucleus, in three patients with essential tremor.Main results: Brief, 15 s stimulations of the target at 10% duty cycle with low-intensity ultrasound, repeated less than 30 times over a period of 90 min, nearly abolished tremor (98% and 97% tremor amplitude reduction) in 2 out of 3 patients. The effect was observed within seconds of the stimulation onset and increased with ultrasound exposure time. The effect gradually vanished following the stimulation, suggesting that the stimulation was safe with no harmful long-term consequences detected.Significance: This result demonstrates that low-intensity focused ultrasound can robustly modulate deep brain regions in humans with notable effects on overt motor behavior.

Device for Multifocal Delivery of Ultrasound Into Deep Brain Regions in Humans

Low-intensity focused ultrasound provides the means to noninvasively stimulate or release drugs in specified deep brain targets. However, successful clinical translations require hardware that maximizes acoustic transmission through the skull, enables flexible electronic steering, and provides accurate and reproducible targeting while minimizing the use of MRI. We have developed a device that addresses these practical requirements. The device delivers ultrasound through the temporal and parietal skull windows, which minimize the attenuation and distortions of the ultrasound by the skull. The device consists of 252 independently controlled elements, which provides the ability to modulate multiple deep brain targets at a high spatiotemporal resolution, without the need to move the device or the subject. And finally, the device uses a mechanical registration method that enables accurate deep brain targeting both inside and outside of the MRI. Using this method, a single MRI scan is necessary for accurate targeting; repeated subsequent treatments can be performed reproducibly in an MRI-free manner. We validated these functions by transiently modulating specific deep brain regions in two patients with treatment-resistant depression.

Stem cell-derived brain organoids for controlled studies of transcranial neuromodulation

Transcranial neuromodulation methods have the potential to diagnose and treat brain disorders at their neural source in a personalized manner. However, it has been difficult to investigate the direct effects of transcranial neuromodulation on neurons in human brain tissue. Here, we show that human brain organoids provide a detailed and artifact-free window into neuromodulation-evoked electrophysiological effects. We derived human cortical organoids from induced pluripotent stem cells and implanted 32-channel electrode arrays. Each organoid was positioned in the center of the human skull and subjected to low-intensity transcranial focused ultrasound. We found that ultrasonic stimuli modulated network activity in the gamma and delta ranges of the frequency spectrum. The effects on the neural networks were a function of the ultrasound stimulation frequency. High gamma activity remained elevated for at least 20 minutes following stimulation offset. This approach is expected to provide controlled studies of the effects of ultrasound and other transcranial neuromodulation modalities on human brain tissue.

The development of microfabricated solenoids with magnetic cores for micromagnetic neural stimulation

Electrical stimulation via invasive microelectrodes is commonly used to treat a wide range of neurological and psychiatric conditions. Despite its remarkable success, the stimulation performance is not sustainable since the electrodes become encapsulated by gliosis due to foreign body reactions. Magnetic stimulation overcomes these limitations by eliminating the need for a metal-electrode contact. Here, we demonstrate a novel microfabricated solenoid inductor (80 µm × 40 µm) with a magnetic core that can activate neuronal tissue. The characterization and proof-of-concept of the device raise the possibility that micromagnetic stimulation solenoids that are small enough to be implanted within the brain may prove to be an effective alternative to existing electrode-based stimulation devices for chronic neural interfacing applications.

Spectral Features Based Decoding of Task Engagement: The Role of Theta and High Gamma Bands in Cognitive Control

This paper analyzes local field potentials (LFP) from 10 human subjects to discover frequency-dependent biomarkers of cognitive conflict. We utilize cortical and sub-cortical LFP recordings from the subjects during a cognitive task known as the Multi-Source Interference Task (MSIT). We decode the task engagement and discover biomarkers that may facilitate closed-loop neuromodulation to enhance cognitive control. First, we show that spectral power features in predefined frequency bands can be used to classify task and non-task segments with a median accuracy of 88.1%. Here the features are first ranked using the Bayes Factor and then used as inputs to subject-specific linear support vector machine classifiers. Second, we show that theta (4-8 Hz) band, and high gamma (65-200 Hz) band oscillations are modulated during the task performance. Third, by isolating time-series from specific brain regions of interest, we observe that a subset of the dorsolateral prefrontal cortex features is sufficient to decode the task states. The paper shows that cognitive control evokes robust neurological signatures, especially in the prefrontal cortex (PFC).

Decoding Human Cognitive Control Using Functional Connectivity of Local Field Potentials

Many patients with mental illnesses characterized by impaired cognitive control have no relief from gold-standard clinical treatments resulting in a pressing need for new alternatives. This paper develops a neural decoder to detect task engagement in ten human subjects during a conflict-based behavioral task known as the multi-source interference task (MSIT). Task engagement is of particular interest here because closed-loop brain stimulation during those states can augment decision-making. The functional connectivity patterns of the electrodes are extracted. A principal component analysis of these patterns is carried out and the ranked principal components are used as inputs to train subject-specific linear support vector machine classifiers. In this paper, we show that task engagement can be differentiated from background brain activity with a median accuracy of 89.7%. This was accomplished by constructing distributed functional networks from local field potentials recording during the task performance. A further challenge is that goal-directed efforts take place over higher temporal resolution. Task engagement must thus be detected at a similar rate for proactive intervention. We show that our algorithms can detect task engagement from neural recordings in less than 2 seconds; this can be further improved using an application-specific device.

Deep Brain Stimulation for Obsessive-Compulsive Disorder: Real World Experience Post-FDA-Humanitarian Use Device Approval

Background: While case series have established the efficacy of deep brain stimulation (DBS) in treating obsessive-compulsive disorder (OCD), it has been our experience that few OCD patients present without comorbidities that affect outcomes associated with DBS treatment. Here we present our experience with DBS therapy for OCD in patients who all have comorbid disease, together with the results of our programming strategies. Methods: For this case series, we assessed five patients who underwent ventral capsule/ventral striatum (VC/VS) DBS for OCD between 2015 and 2019 at the University of Colorado Hospital. Every patient in this cohort exhibited comorbidities, including substance use disorders, eating disorder, tic disorder, and autism spectrum disorder. We conducted an IRB-approved, retrospective study of programming modifications and treatment response over the course of DBS therapy. Results: In addition to patients' subjective reports of improvement, we observed significant improvement in the Yale-Brown Obsessive-Compulsive Scale (44%), the Montgomery-Asberg Depression Rating Scale (53%), the Quality of Life Enjoyment and Satisfaction Questionnaire (27%), and the Hamilton Anxiety Rating scales (34.9%) following DBS. With respect to co-morbid disease, there was a significant improvement in a patient with tic disorder's Total Tic Severity Score (TTSS) (p = 0.005). Conclusions: DBS remains an efficacious tool for the treatment of OCD, even in patients with significant comorbidities in whom DBS has not previously been investigated. Efficacious treatment results not only from the accurate placement of the electrodes by the surgeon but also from programming by the psychiatrist.

Parallel transmission to reduce absorbed power around deep brain stimulation devices in MRI: Impact of number and arrangement of transmit channels

PURPOSE

To assess the mean and variance performance of parallel transmission (pTx) coils for reduction of the absorbed power around electrodes (APAE) in patients implanted with deep brain stimulation (DBS) devices.

METHODS

We simulated 4 pTx coils (8 and 16 channels, head and body coils) and a birdcage body coil. We characterized the RF safety risk using the APAE, which is the integral of the deposited power (in Watts) in a small cylindrical volume of brain tissue surrounding the electrode tips. We assessed the APAE mean and variance by simulation of 5 realistic DBS patient models that include the full DBS implant length, extracranial loops, and implanted pulse generator.

RESULTS

PTx coils with 8 (16) channels were able to reduce the APAE by >18× (>169×) compared to the birdcage coil in average for all patient models, at no cost in term of flip angle uniformity or global specific absorption rate (SAR). Moreover, local pTx coils performed significantly better than body arrays.

CONCLUSION

PTx is a possible solution to the problem of RF heating of DBS patients when performing MRI, but the large interpatient variability of the APAE indicates that patient-specific safety monitoring may be needed.

Decoding task engagement from distributed network electrophysiology in humans

OBJECTIVE

Here, our objective was to develop a binary decoder to detect task engagement in humans during two distinct, conflict-based behavioral tasks. Effortful, goal-directed decision-making requires the coordinated action of multiple cognitive processes, including attention, working memory and action selection. That type of mental effort is often dysfunctional in mental disorders, e.g. when a patient attempts to overcome a depression or anxiety-driven habit but feels unable. If the onset of engagement in this type of focused mental activity could be reliably detected, decisional function might be augmented, e.g. through neurostimulation. However, there are no known algorithms for detecting task engagement with rapid time resolution.

APPROACH

We defined a new network measure, fixed canonical correlation (FCCA), specifically suited for neural decoding applications. We extracted FCCA features from local field potential recordings in human volunteers to give a temporally continuous estimate of mental effort, defined by engagement in experimental conflict tasks.

MAIN RESULTS

Using a small number of features per participant, we accurately decoded and distinguished task engagement from other mental activities. Further, the decoder distinguished between engagement in two different conflict-based tasks within seconds of their onset.

SIGNIFICANCE

These results demonstrate that network-level brain activity can detect specific types of mental efforts. This could form the basis of a responsive intervention strategy for decision-making deficits.

Consistent linear and non-linear responses to invasive electrical brain stimulation across individuals and primate species with implanted electrodes

BACKGROUND

Electrical neuromodulation via implanted electrodes is used in treating numerous neurological disorders, yet our knowledge of how different brain regions respond to varying stimulation parameters is sparse.

OBJECTIVE/HYPOTHESIS

We hypothesized that the neural response to electrical stimulation is both region-specific and non-linearly related to amplitude and frequency.

METHODS

We examined evoked neural responses following 400 ms trains of 10-400 Hz electrical stimulation ranging from 0.1 to 10 mA. We stimulated electrodes implanted in cingulate cortex (dorsal anterior cingulate and rostral anterior cingulate) and subcortical regions (nucleus accumbens, amygdala) of non-human primates (NHP, N = 4) and patients with intractable epilepsy (N = 15) being monitored via intracranial electrodes. Recordings were performed in prefrontal, subcortical, and temporal lobe locations.

RESULTS

In subcortical regions as well as dorsal and rostral anterior cingulate cortex, response waveforms depended non-linearly on frequency (Pearson's linear correlation r < 0.39), but linearly on current (r > 0.58). These relationships between location, and input-output characteristics were similar in homologous brain regions with average Pearson's linear correlation values r > 0.75 between species and linear correlation values between participants r > 0.75 across frequency and current values per brain region. Evoked waveforms could be described by three main principal components (PCs) which allowed us to successfully predict response waveforms across individuals and across frequencies using PC strengths as functions of current and frequency using brain region specific regression models.

CONCLUSIONS

These results provide a framework for creation of an atlas of input-output relationships which could be used in the principled selection of stimulation parameters per brain region.

Animal Models and Animal Experimentation in the Development of Deep Brain Stimulation: From a Specific Controversy to a Multidimensional Debate

In this article, we explore a specific controversy about animal experimentation and animal models in the recent history of deep brain stimulation (DBS), and we question its ramifications. DBS development intertwines clinical practice with fundamental research and stands at the crossroads of multiple legacies. We take up the various issues and controversies embedded in this rarely addressed dispute, from a standpoint that combines socio-anthropological and legal aspects. Our starting point is a debate on the role of animal experimentation in the development of DBS between Jarrod Bailey, a researcher promoting the abolition of animal experimentation, and Alim Louis Benabid, Marwan Hariz, and Mahlon DeLong, three key figures in the area of DBS and neuroscience. By clarifying the positions of the different protagonists and retracing the issues raised in these discussions, our objective is to show how this specific debate has extended from its initial space and how it provides an object of study with heuristic scope. We first present this partially polemic discussion about the history of DBS, and its link with a more general debate on the validity and use of animal models and the need for animal experiments. Then, we raise the issue of the relations and interactions between experiments on animals and on humans in the logics of biomedical innovation. The third step is to situate the discussion within the wider framework of opposition towards animal experimentation and the promotion of animal' rights. Finally, combining these interweaved issues, possible implications emerge regarding the future of DBS. We show that behind these several controversies lie the question of translational research and the model of medicine upheld by DBS. We describe how the technology contributes to blurring the lines between research (fundamental, preclinical and clinical research) and care, as well as between humans and animals as substrates and objects of knowledge. The dynamics of DBS future development might then become a point of convergence for neuroscientists and animal rights defenders' interests.

The 'virtual DBS population': five realistic computational models of deep brain stimulation patients for electromagnetic MR safety studies

We design, develop, and disseminate a 'virtual population' of five realistic computational models of deep brain stimulation (DBS) patients for electromagnetic (EM) analysis. We found five DBS patients in our institution' research patient database who received high quality post-DBS surgery computer tomography (CT) examinations of the head and neck. Three patients have a single implanted pulse generator (IPG) and the two others have two IPGs (one for each lead). Moreover, one patient has two abandoned leads on each side of the head. For each patient, we combined the head and neck volumes into a 'virtual CT', from which we extracted the full-length DBS path including the IPG, extension cables, and leads. We corrected topology errors in this path, such as self-intersections, using a previously published optimization procedure. We segmented the virtual CT volume into bones, internal air, and soft tissue classes and created two-manifold, watertight surface meshes of these distributions. In addition, we added a segmented model of the brain (grey matter, white matter, eyes and cerebrospinal fluid) to one of the model (nickname Freddie) that was derived from a T1-weighted MR image obtained prior to the DBS implantation. We simulated the EM fields and specific absorption rate (SAR) induced at 3 Tesla by a quadrature birdcage body coil in each of the five patient models using a co-simulation strategy. We found that inter-subject peak SAR variability across models was independent of the target averaging mass and equal to ~45%. In our simulations of the full brain segmentation and six simplified versions of the Freddie model, the error associated with incorrect dielectric property assignment around the DBS electrodes was greater than the error associated with modeling the whole model as a single tissue class. Our DBS patient models are freely available on our lab website (Webpage of the Martinos Center Phantom Resource 2018 https://phantoms.martinos.org/Main_Page).

Altering alpha-frequency brain oscillations with rapid analog feedback-driven neurostimulation

Oscillations of the brain's local field potential (LFP) may coordinate neural ensembles and brain networks. It has been difficult to causally test this model or to translate its implications into treatments, because there are few reliable ways to alter LFP oscillations. We developed a closed-loop analog circuit to enhance brain oscillations by feeding them back into cortex through phase-locked transcranial electrical stimulation. We tested the system in a rhesus macaque with chronically implanted electrode arrays, targeting 8-15 Hz (alpha) oscillations. Ten seconds of stimulation increased alpha oscillatory power for up to 1 second after stimulation offset. In contrast, open-loop stimulation decreased alpha power. There was no effect in the neighboring 15-30 Hz (beta) LFP rhythm or on a neighboring array that did not participate in closed-loop feedback. Analog closed-loop neurostimulation might thus be a useful strategy for altering brain oscillations, both for basic research and the treatment of neuro-psychiatric disease.

Neuroscientifically Informed Formulation and Treatment Planning for Patients With Obsessive-Compulsive Disorder: A Review

IMPORTANCE

Obsessive-compulsive disorder (OCD) is a common and often debilitating psychiatric illness. Recent advances in the understanding of the neuroscience of OCD have provided valuable insights that have begun to transform the way we think about the management of this disorder. This educational review provides an integrated neuroscience perspective on formulation and treatment planning for patients with OCD. The article is organized around key neuroscience themes most relevant for OCD.

OBSERVATIONS

An integrated neuroscience formulation of OCD is predicated on a fundamental understanding of phenomenology and symptom dimensions, fear conditioning and extinction, neurochemistry, genetics and animal models, as well as neurocircuitry and neurotherapeutics. Symptom dimensions provide a means to better understand the phenotypic heterogeneity within OCD with an eye toward more personalized treatments. The concept of abnormal fear extinction is central to OCD and to the underlying therapeutic mechanism of exposure and response prevention. A framework for understanding the neurochemistry of OCD focuses on both traditional monoaminergic systems and more recent evidence of glutamatergic and γ-aminobutyric acid-ergic dysfunction. Obsessive-compulsive disorder is highly heritable, and future work is needed to understand the contribution of genes to underlying pathophysiology. A circuit dysregulation framework focuses on cortico-striato-thalamo-cortical circuit dysfunction and the development of neurotherapeutic approaches targeting this circuit. The impact of these concepts on how we think about OCD diagnosis and treatment is discussed. Suggestions for future investigations that have the potential to further enhance the clinical management of OCD are presented.

CONCLUSIONS AND RELEVANCE

These key neuroscience themes collectively inform formulation and treatment planning for patients with OCD. The ultimate goal is to increase crosstalk between clinicians and researchers in an effort to facilitate translation of advances in neuroscience research to improved care for patients with OCD.

Deep Brain Stimulation in Psychiatry: Mechanisms, Models, and Next-Generation Therapies

Deep brain stimulation has preliminary evidence of clinical efficacy, but has been difficult to develop into a robust therapy, in part because its mechanisms are incompletely understood. We review evidence from movement and psychiatric disorder studies, with an emphasis on how deep brain stimulation changes brain networks. From this, we argue for a network-oriented approach to future deep brain stimulation studies. That network approach requires methods for identifying patients with specific circuit/network deficits. We describe how dimensional approaches to diagnoses may aid that identification. We discuss the use of network/circuit biomarkers to develop self-adjusting "closed loop" systems.

Realistic modeling of deep brain stimulation implants for electromagnetic MRI safety studies

We propose a framework for electromagnetic (EM) simulation of deep brain stimulation (DBS) patients in radiofrequency (RF) coils. We generated a model of a DBS patient using post-operative head and neck computed tomography (CT) images stitched together into a 'virtual CT' image covering the entire length of the implant. The body was modeled as homogeneous. The implant path extracted from the CT data contained self-intersections, which we corrected automatically using an optimization procedure. Using the CT-derived DBS path, we built a model of the implant including electrodes, helicoidal internal conductor wires, loops, extension cables, and the implanted pulse generator. We also built four simplified models with straight wires, no extension cables and no loops to assess the impact of these simplifications on safety predictions. We simulated EM fields induced by the RF birdcage body coil in the body model, including at the DBS lead tip at both 1.5 Tesla (64 MHz) and 3 Tesla (123 MHz). We also assessed the robustness of our simulation results by systematically varying the EM properties of the body model and the position and length of the DBS implant (sensitivity analysis). The topology correction algorithm corrected all self-intersection and curvature violations of the initial path while introducing minimal deformations (open-source code available at http://ptx.martinos.org/index.php/Main_Page). The unaveraged lead-tip peak SAR predicted by the five DBS models (0.1 mm resolution grid) ranged from 12.8 kW kg-1 (full model, helicoidal conductors) to 43.6 kW kg-1 (no loops, straight conductors) at 1.5 T (3.4-fold variation) and 18.6 kW kg-1 (full model, straight conductors) to 73.8 kW kg-1 (no loops, straight conductors) at 3 T (4.0-fold variation). At 1.5 T and 3 T, the variability of lead-tip peak SAR with respect to the conductivity ranged between 18% and 30%. Variability with respect to the position and length of the DBS implant ranged between 9.5% and 27.6%.

Towards miniaturized closed-loop optogenetic stimulation devices

OBJECTIVE

Electrical brain stimulation provides therapeutic benefits for patients with drug-resistant neurological disorders. It, however, has restricted access to cell-type selectivity which limits its treatment effectiveness. Optogenetics, in contrast, enables precise targeting of a specific cell type which can address the issue with electrical brain stimulation. It, nonetheless, disregards real-time brain responses in delivering optimized stimulation to target cells. Closed-loop optogenetics, on the other hand, senses the difference between normal and abnormal states of the brain, and modulates stimulation parameters to achieve the desired stimulation outcome. Current review articles on closed-loop optogenetics have focused on its theoretical aspects and potential benefits. A review of the recent progress in miniaturized closed-loop optogenetic stimulation devices is thus needed.

APPROACH

This paper presents a comprehensive study on the existing miniaturized closed-loop optogenetic stimulation devices and their internal components.

MAIN RESULTS

Hardware components of closed-loop optogenetic stimulation devices including electrode, light-guiding mechanism, optical source, neural recorder, and optical stimulator are discussed. Next, software modules of closed-loop optogenetic stimulation devices including feature extraction, classification, control, and stimulation parameter modulation are described. Then, the existing devices are categorized into open-loop and closed-loop groups, and the combined operation of their neural recorder, optical stimulator, and control approach is discussed. Finally, the challenges in the design and implementation of closed-loop optogenetic stimulation devices are presented, suggestions on how to tackle these challenges are given, and future directions for closed-loop optogenetics are stated.

SIGNIFICANCE

A generic architecture for closed-loop optogenetic stimulation devices involving both hardware and software perspectives is devised. A comprehensive investigation into the most current miniaturized and tetherless closed-loop optogenetic stimulation devices is given. A detailed comparison of the closed-loop optogenetic stimulation devices is presented.

Deep brain stimulation of the anterior limb of the internal capsule for treatment of therapy-refractory obsessive compulsive disorder (OCD): a case study highlighting neurocognitive and psychiatric changes

Obsessive compulsive disorder (OCD) is an anxiety disorder characterized by repeated, unwanted thoughts and behaviors. Individuals with this condition often experience significant emotional distress secondary to their symptoms. Additionally, impairments in attention/concentration, processing speed, and executive functions are typically observed. The exact pathology of OCD remains unknown; consequently, it can be difficult to treat patients with severe symptomatology. Deep brain stimulation (DBS) may be a viable treatment option for individuals who do not respond to medication and/or cognitive behavioral therapy. The following case discusses DBS of the anterior limb of the internal capsule for a patient with severe, therapy-refractory OCD, including pre- to postoperative neurocognitive and psychiatric changes.

Closed-loop neuromodulation systems: next-generation treatments for psychiatric illness

Despite deep brain stimulation's positive early results in psychiatric disorders, well-designed clinical trials have yielded inconsistent clinical outcomes. One path to more reliable benefit is closed-loop therapy: stimulation that is automatically adjusted by a device or algorithm in response to changes in the patient's electrical brain activity. These interventions may provide more precise and patient-specific treatments. This article first introduces the available closed-loop neuromodulation platforms, which have shown clinical efficacy in epilepsy and strong early results in movement disorders. It discusses the strengths and limitations of these devices in the context of psychiatric illness. It then describes emerging technologies to address these limitations, including pre-clinical developments such as wireless deep neurostimulation and genetically targeted neuromodulation. Finally, ongoing challenges and limitations for closed-loop psychiatric brain stimulation development, most notably the difficulty of identifying meaningful biomarkers for titration, are discussed. This is considered in the recently-released Research Domain Criteria (RDoC) framework, and how neuromodulation and RDoC are jointly very well suited to address the problem of treatment-resistant illness is described.

Predicting local field potentials with recurrent neural networks

We present a Recurrent Neural Network using LSTM (Long Short Term Memory) that is capable of modeling and predicting Local Field Potentials. We train and test the network on real data recorded from epilepsy patients. We construct networks that predict multi-channel LFPs for 1, 10, and 100 milliseconds forward in time. Our results show that prediction using LSTM outperforms regression when predicting 10 and 100 millisecond forward in time.