Closing the loop in psychiatric deep brain stimulation: physiology, psychometrics, and plasticity

Failure modes and mitigations for Bayesian optimization of neuromodulation parameters

Objective.Precision medicine holds substantial promise for tailoring neuromodulation techniques to the symptomatology of individual patients. Precise selection of stimulation parameters for individual patients requires the development of robust optimization techniques. However, standard optimization approaches, like Bayesian optimization, have historically been assessed and developed for applications with far less noise than is typical in neuro-psychiatric outcome measures and with minimal focus on parameter safety.Approach.We conducted a literature review of individual effects in neurological and psychiatric applications to build a series of simulated patient responses of varying signal to noise ratio. Using these simulations, we assessed whether existing standards in Bayesian optimization are sufficient for robustly optimizing such effects.Main results.For effect sizes below a Cohen's d of 0.3, standard Bayesian optimization methods failed to consistently identify optimal parameters. This failure primarily results from over-sampling of the boundaries of the space as the number of samples increases, because the variance on the edges becomes disproportionately greater than in the remainder of parameter space. Using a combination of an input warp and a boundary avoiding Iterated Brownian-bridge kernel we demonstrated robust Bayesian optimization performance for problems with a Cohen's d effect size as low as 0.1.Significance.Our results demonstrate that caution should be taken when applying standard Bayesian optimization in neuromodulation applications with potentially low effect sizes, as standard algorithms are at high risk of converging to local rather than global optima. Mitigating techniques, like boundary avoidance, are effective and should be used to improve robustness.

Challenges and opportunities of acquiring cortical recordings for chronic adaptive deep brain stimulation

Deep brain stimulation (DBS), a proven treatment for movement disorders, also holds promise for the treatment of psychiatric and cognitive conditions. However, for DBS to be clinically effective, it may require DBS technology that can alter or trigger stimulation in response to changes in biomarkers sensed from the patient's brain. A growing body of evidence suggests that such adaptive DBS is feasible, it might achieve clinical effects that are not possible with standard continuous DBS and that some of the best biomarkers are signals from the cerebral cortex. Yet capturing those markers requires the placement of cortex-optimized electrodes in addition to standard electrodes for DBS. In this Perspective we argue that the need for cortical biomarkers in adaptive DBS and the unfortunate convergence of regulatory and financial factors underpinning the unavailability of cortical electrodes for chronic uses threatens to slow down or stall research on adaptive DBS and propose public-private partnerships as a potential solution to such a critical technological gap.

Erasing "bad memories": reversing aberrant synaptic plasticity as therapy for neurological and psychiatric disorders

Dopamine modulates corticostriatal plasticity in both the direct and indirect pathways of the cortico-striato-thalamo-cortical (CSTC) loops. These gradual changes in corticostriatal synaptic strengths produce long-lasting changes in behavioral responses. Under normal conditions, these mechanisms enable the selection of the most appropriate responses while inhibiting others. However, under dysregulated dopamine conditions, including a lack of dopamine release or dopamine signaling, these mechanisms could lead to the selection of maladaptive responses and/or the inhibition of appropriate responses in an experience-dependent and task-specific manner. In this review, we propose that preventing or reversing such maladaptive synaptic strengths and erasing such aberrant "memories" could be a disease-modifying therapeutic strategy for many neurological and psychiatric disorders. We review evidence from Parkinson's disease, drug-induced parkinsonism, L-DOPA-induced dyskinesia, obsessive-compulsive disorder, substance use disorders, and depression as well as research findings on animal disease models. Altogether, these studies allude to an emerging theme in translational neuroscience and promising new directions for therapy development. Specifically, we propose that combining pharmacotherapy with behavioral therapy or with deep brain stimulation (DBS) could potentially cause desired changes in specific neural circuits. If successful, one important advantage of correcting aberrant synaptic plasticity is long-lasting therapeutic effects even after treatment has ended. We will also discuss the potential molecular targets for these therapeutic approaches, including the cAMP pathway, proteins involved in synaptic plasticity as well as pathways involved in new protein synthesis. We place special emphasis on RNA binding proteins and epitranscriptomic mechanisms, as they represent a new frontier with the distinct advantage of rapidly and simultaneously altering the synthesis of many proteins locally.

Neurostimulation to Improve Cognitive Flexibility

Cognitive flexibility, the capacity to adapt behaviors in response to changing environments, is impaired across mental illnesses, including depression, anxiety, addiction, and obsessive-compulsive disorder. Cortico-striatal-cortical circuits are integral to cognition and goal-directed behavior and disruptions in these circuits are linked to cognitive inflexibility in mental illnesses. We review evidence that neurostimulation of these circuits can improve cognitive flexibility and ameliorate symptoms, and that this may be a mechanism of action of current clinical therapies. Further, we discuss how animal models can offer insights into the mechanisms underlying cognitive flexibility and effects of neurostimulation. We review research from animal studies that may, if translated, yield better approaches to modulating flexibility. Future research should focus on refining definitions of cognitive flexibility, improving detection of impaired flexibility, and developing new methods for optimizing neurostimulation parameters. This could enhance neurostimulation therapies through more personalized treatments that leverage cognitive flexibility to improve patient outcomes.

Novel NIBS in psychiatry: Unveiling TUS and TI for research and treatment

Mental disorders pose a significant global burden and constitute a major cause of disability worldwide. Despite strides in treatment, a substantial number of patients do not respond adequately, underscoring the urgency for innovative approaches. Traditional non-invasive brain stimulation techniques show promise, yet grapple with challenges regarding efficacy and specificity. Variations in mechanistic understanding and reliability among non-invasive brain stimulation methods are common, with limited spatial precision and physical constraints hindering the ability to target subcortical areas often implicated in the disease aetiology. Novel techniques such as transcranial ultrasonic stimulation and temporal interference stimulation have gained notable momentum in recent years, possibly addressing these shortcomings. Transcranial ultrasonic stimulation (TUS) offers exceptional spatial precision and deeper penetration compared with conventional electrical and magnetic stimulation techniques. Studies targeting a diverse array of brain regions have shown its potential to affect neuronal excitability, functional connectivity and symptoms of psychiatric disorders such as major depressive disorder. Nevertheless, challenges such as target planning and addressing acoustic interactions with the skull must be tackled for its widespread adoption in research and potentially clinical settings. Similar to transcranial ultrasonic stimulation, temporal interference (TI) stimulation offers the potential to target deeper subcortical areas compared with traditional non-invasive brain stimulation, albeit requiring a comparatively higher current for equivalent neural effects. Promising yet still sparse research highlights TI's potential to selectively modulate neuronal activity, showing potential for its utility in psychiatry. Overall, recent strides in non-invasive brain stimulation methods like transcranial ultrasonic stimulation and temporal interference stimulation not only open new research avenues but also hold potential as effective treatments in psychiatry. However, realising their full potential necessitates addressing practical challenges and optimising their application effectively.

Striatal stimulation enhances cognitive control and evidence processing in rodents and humans

Brain disorders, in particular mental disorders, might be effectively treated by direct electrical brain stimulation, but clinical progress requires understanding of therapeutic mechanisms. Animal models have not helped, because there are no direct animal models of mental illness. Here, we propose a potential path past this roadblock, by leveraging a common ingredient of most mental disorders: impaired cognitive control. We previously showed that deep brain stimulation (DBS) improves cognitive control in humans. We now reverse translate that result using a set-shifting task in rats. DBS-like stimulation of the midstriatum improved reaction times without affecting accuracy, mirroring our human findings. Impulsivity, motivation, locomotor, and learning effects were ruled out through companion tasks and model-based analyses. To identify the specific cognitive processes affected, we applied reinforcement learning drift-diffusion modeling. This approach revealed that DBS-like stimulation enhanced evidence accumulation rates and lowered decision thresholds, improving domain-general cognitive control. Reanalysis of prior human data showed that the same mechanism applies in humans. This reverse/forward translational model could have near-term implications for clinical DBS practice and future trial design.

Neural decoding and feature selection methods for closed-loop control of avoidance behavior

Objective.Many psychiatric disorders involve excessive avoidant or defensive behavior, such as avoidance in anxiety and trauma disorders or defensive rituals in obsessive-compulsive disorders. Developing algorithms to predict these behaviors from local field potentials (LFPs) could serve as the foundational technology for closed-loop control of such disorders. A significant challenge is identifying the LFP features that encode these defensive behaviors.Approach.We analyzed LFP signals from the infralimbic cortex and basolateral amygdala of rats undergoing tone-shock conditioning and extinction, standard for investigating defensive behaviors. We utilized a comprehensive set of neuro-markers across spectral, temporal, and connectivity domains, employing SHapley Additive exPlanations for feature importance evaluation within Light Gradient-Boosting Machine models. Our goal was to decode three commonly studied avoidance/defensive behaviors: freezing, bar-press suppression, and motion (accelerometry), examining the impact of different features on decoding performance.Main results.Band power and band power ratio between channels emerged as optimal features across sessions. High-gamma (80-150 Hz) power, power ratios, and inter-regional correlations were more informative than other bands that are more classically linked to defensive behaviors. Focusing on highly informative features enhanced performance. Across 4 recording sessions with 16 subjects, we achieved an average coefficient of determination of 0.5357 and 0.3476, and Pearson correlation coefficients of 0.7579 and 0.6092 for accelerometry jerk and bar press rate, respectively. Utilizing only the most informative features revealed differential encoding between accelerometry and bar press rate, with the former primarily through local spectral power and the latter via inter-regional connectivity. Our methodology demonstrated remarkably low training/inference time and memory usage, requiring<310 ms for training,<0.051 ms for inference, and 16.6 kB of memory, using a single core of AMD Ryzen Threadripper PRO 5995WX CPU.Significance.Our results demonstrate the feasibility of accurately decoding defensive behaviors with minimal latency, using LFP features from neural circuits strongly linked to these behaviors. This methodology holds promise for real-time decoding to identify physiological targets in closed-loop psychiatric neuromodulation.

Optimizing real-time phase detection in diverse rhythmic biological signals for phase-specific neuromodulation

Closed-loop, phase-specific neurostimulation is a powerful method to modulate ongoing brain activity for clinical and research applications. Phase-specific stimulation relies on estimating the phase of an ongoing oscillation in real time and issuing a control command at a target phase. Phase detection algorithms based on Fast Fourier transform (FFT) are widely used due to their computational efficiency and robustness. However, it is unclear how algorithm performance depends on the spectral properties of the input signal and how algorithm parameters can be optimized. We used offline simulation to evaluate the performance of three algorithms (endpoint-corrected Hilbert Transform, Hilbert Transform and phase mapping) on three rhythmic biological signals with distinct spectral properties (rodent hippocampal theta potential, human EEG alpha and human essential tremor). First, we found that algorithm performance was more strongly influenced by signal amplitude and frequency variation compared with signal to noise ratio. Second, our simulations showed that the size of the data window for phase estimation was critical for the performance of FFT-based algorithms, where the optimal data window corresponds to the period of the oscillation. We validated this prediction with real time phase detection of hippocampal theta oscillations in freely behaving rats performing spatial navigation. Our findings define the relationship between signal properties and algorithm performance and provide a convenient method for optimizing FFT-based phase detection algorithms.

Method for Synthetic Generation of LFP Data for Testing of Feature Extraction Algorithms. ANNUAL INTERNATIONAL CONFERENCE OF THE IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY. IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY. ANNUAL INTERNATIONAL CONFERENCE 2024;2024:1-4. [PMID: 40040065 DOI: 10.1109/embc53108.2024.10781522] [Citation(s) in RCA: 0] [Impact Index Per Article: 0]

Recent interest in closed-loop neuromodulation devices has driven development of algorithms capable of real-time biomarker extraction. Synthetic data for tuning algorithmic parameters in various oscillatory cases is a useful tool but must be generated to model realistic neural behavior. We extracted key oscillatory behaviors from rodent LFPs and used this information to create a realistic generation method for synthetic signal production. We then used the generated signals to optimize the feature extraction performance of a real-time feature extraction algorithm. The results of the algorithm testing closely mirrored results from testing on recorded neural LFPs and resembled this real data more closely than a simplistic model of synthetic neural data.

Concurrent inflammation-related brain reorganization in multiple sclerosis and depression

BACKGROUND

Neuroinflammation affects brain tissue integrity in multiple sclerosis (MS) and may have a role in major depressive disorder (MDD). Whether advanced magnetic resonance imaging characteristics of the gray-to-white matter border serve as proxy of neuroinflammatory activity in MDD and MS remain unknown.

METHODS

We included 684 participants (132 MDD patients with recurrent depressive episodes (RDE), 70 MDD patients with a single depressive episode (SDE), 222 MS patients without depressive symptoms (nMS), 58 MS patients with depressive symptoms (dMS), and 202 healthy controls (HC)). 3 T-T1w MRI-derived gray-to-white matter contrast (GWc) was used to reconstruct and characterize connectivity alterations of GWc-covariance networks by means of modularity, clustering coefficient, and degree. A cross-validated support vector machine was used to test the ability of GWc to stratify groups according to their depression symptoms, measured with BDI, at the single-subject level in MS and MDD independently.

FINDINGS

MS and MDD patients showed increased modularity (ANOVA partial-η2 = 0.3) and clustering (partial-η2 = 0.1) compared to HC. In the subgroups, a linear trend analysis attested a gradient of modularity increases in the form: HC, dMS, nMS, SDE, and RDE (ANOVA partial-η2 = 0.28, p < 0.001) while this trend was less evident for clustering coefficient. Reduced morphological integrity (GWc) was seen in patients with increased depressive symptoms (partial-η2 = 0.42, P < 0.001) and was associated with depression scores across patient groups (r = -0.2, P < 0.001). Depressive symptoms in MS were robustly classified (88 %).

CONCLUSIONS

Similar structural network alterations in MDD and MS exist, suggesting possible common inflammatory events like demyelination, neuroinflammation that are caught by GWc analyses. These alterations may vary depending on the severity of symptoms and in the case of MS may elucidate the occurrence of comorbid depression.

Individualized Closed-Loop Acoustic Stimulation Suggests an Alpha Phase Dependence of Sound Evoked and Induced Brain Activity Measured with EEG Recordings

Following repetitive visual stimulation, post hoc phase analysis finds that visually evoked response magnitudes vary with the cortical alpha oscillation phase that temporally coincides with sensory stimulus. This approach has not successfully revealed an alpha phase dependence for auditory evoked or induced responses. Here, we test the feasibility of tracking alpha with scalp electroencephalogram (EEG) recordings and play sounds phase-locked to individualized alpha phases in real-time using a novel end-point corrected Hilbert transform (ecHT) algorithm implemented on a research device. Based on prior work, we hypothesize that sound-evoked and induced responses vary with the alpha phase at sound onset and the alpha phase that coincides with the early sound-evoked response potential (ERP) measured with EEG. Thus, we use each subject's individualized alpha frequency (IAF) and individual auditory ERP latency to define target trough and peak alpha phases that allow an early component of the auditory ERP to align to the estimated poststimulus peak and trough phases, respectively. With this closed-loop and individualized approach, we find opposing alpha phase-dependent effects on the auditory ERP and alpha oscillations that follow stimulus onset. Trough and peak phase-locked sounds result in distinct evoked and induced post-stimulus alpha level and frequency modulations. Though additional studies are needed to localize the sources underlying these phase-dependent effects, these results suggest a general principle for alpha phase-dependence of sensory processing that includes the auditory system. Moreover, this study demonstrates the feasibility of using individualized neurophysiological indices to deliver automated, closed-loop, phase-locked auditory stimulation.

Brain Pioneers and Moral Entanglement: An Argument for Post-trial Responsibilities in Neural-Device Trials

We argue that in implanted neurotechnology research, participants and researchers experience what Henry Richardson has called "moral entanglement." Participants partially entrust researchers with access to their brains and thus to information that would otherwise be private, leading to created intimacies and special obligations of beneficence for researchers and research funding agencies. One of these obligations, we argue, is about continued access to beneficial technology once a trial ends. We make the case for moral entanglement in this context through exploration of participants' vulnerability, uncompensated risks and burdens, depth of relationship with the research team, and dependence on researchers in implanted neurotechnology trials.