The performance of microelectrodes for neural stimulation is controlled by their impedance and their charge-storage capacity. Porous electrode systems such as sputtered iridium oxide film (SIROF) microelectrodes (200-2000 μm2) and ultramicroelectrodes (<200 μm2) exhibit electrochemical properties suitable for neural stimulation.The objective of the present work is to apply the measurement model approach pioneered by the Orazem group to impedance data collected on SIROF microelectrodes in phosphate-buffered saline as a function of applied potential. The data were collected in triplicate, allowing assessment of the stochastic error structure. The interpretation model accounts for the porous electrode structure and the influence of the iridium oxidation state transitions on the impedance response.The information gained from in vitro experiments can guide interpretation of the impedance response for similar electrodes conducted in animal or human studies. While the impedance conducted in vivo is certainly influenced by the different environment, in vitro studies facilitate interpretation of electrode-electrolyte properties, which comprise a component of impedance measurements performed in vivo.The present work demonstrates a philosophy of interpretation modeling that accounts for both the physics and chemistry of the system and the error structure of the measurements. It is common for authors reporting the properties of electrodes used in neuromodulation applications to include EIS measurements. The depth of interpretation of EIS data is often modest, and the checking and handling of measurement error absent. The methodology described in the present work is generally applicable to all electrodes used in electrical-stimulation-based neuromodulation and will provide a framework for understanding in vivo electrode behavior and informing the optimization of stimulation electrode coatings. As such, the present work has broad application to neural engineering.

Engineering replacement organs is the next frontier in therapeutic technologies. Yet, the integration of innervation-critical for organ development, function, and homeostasis-remains underexplored. This review highlights the role of neural inputs in regulating critical organs including pancreas, liver, salivary gland, and spleen. We examine organ-specific neuroanatomy and emerging strategies to incorporate neuronal-axonal networks in engineered organs, drawing from innovations in scaffold design, multi-cell culture techniques, neural engineering, and biofabrication. Finally, we discuss tools for evaluating innervation across in vitro, preclinical, and clinical settings, advocating for innervation as a core design element in next-generation artificial organs.

Introduction Brain‐computer interfaces (BCIs) offer high throughput decoding of neural signals. Current clinical implications include communication and motor rehabilitation. Invasive implants offer high‐fidelity recordings at the cost of craniotomy and higher morbidity. Conversely, non‐invasive modalities suffer from poor spatial resolution. Endovascular BCIs entail minimally invasive procedures via vascular access for cortical activity recording. They have the potential to both decrease morbidity, while still offering high‐fidelity neural encoding. This State‐of‐the‐Art (SotA) review aims to integrate both preclinical and clinical evidence. It seeks to elucidate feasibility, safety, and translational research gaps. Materials and Methods A State‐of‐the‐Art (SotA) methodologic review of Scopus, Web of Science, Ovid MEDLINE, EMBASE, IEEE Xplore, Google Scholar, and ClinicalTrials.gov was conducted. All applicable studies up to August 2025 for endovascular BCIs were screened. These included any devices utilized for recording, stimulation, device design, and/or clinical applications. Aggregate data and quantitative outcomes were extracted, including but not limited to: signal quality, safety, functional results, and/or multivariate factors related to the procedures. Priority was given to preclinical and first‐in‐human studies (NCT03834857 and NCT05035823). Results Preclinical models indicate that stent‐like electrode arrays inserted into cortical veins, including the superior sagittal sinus and tributaries, are chronically stable and capture electrocorticographic signals without inducing parenchymal damage. Early human feasibility trials with the Stentrode™ system indicate accurate decoding of motor intention in patients with severe paralysis, allowing for functional digital communication and device control. Signal fidelity was greater than scalp EEG but smaller than subdural grids, with adaptive machine‐learning decoding improving decoding accuracy. Procedural safety was similar to venous stenting, with thromboembolic complications only and no long‐term parenchymal damage. Advances in miniaturized amplifier technology, wireless telemetry, and biocompatible flexible scaffolds enable long‐term clinical translation. Limitations are sparse electrode density caused by vascular anatomy, unknown long‐term endothelial response, and developing decoding algorithms. Conclusion Endovascular BCIs are a disruptive technology with potentially significant clinical impact. It is an innovative convergence between neural engineering and neurointervention. Although sparse, there is a growing body of early evidence that indicates safety, feasibility, and rehabilitation benefit. However, many challenges remain. Future trials will need to: (1) explore long‐term durability and neointimal integration, (2) enhance decoding accuracy, and (3) establish relative effectiveness when compared to standard BCI modalities. Furthermore, serious efforts must be made to address ethics and regulatory issues. Neurointerventionalists are best poised to standardize procedural protocols and guide translational research efforts.

Abstract The emerging field of neuromorphic computing for edge control applications poses the need to quantitatively estimate and limit the number of spiking neurons, to reduce network complexity and optimize the number of neurons per core and hence, the chip size, in an application-specific neuromorphic hardware. While rate-encoding for spiking neurons provides a robust way to encode signals with the same number of neurons as an ANN, it often lacks precision. To achieve the desired accuracy, a population of neurons is often needed to encode the complete range of input signals. However, using population encoding immensely increases the total number of neurons required for a particular application, thus increasing the power consumption and on-board resource utilization. A transition from two neurons to a population of neurons for the LQR control of a cartpole is shown in this work. The near-linear behavior of a Leaky-Integrate-and-Fire neuron can be exploited to achieve the Linear Quadratic Regulator (LQR) control of a cartpole system. This has been shown in simulation, followed by a demonstration on a single-neuron hardware, known as Lu.i. The improvement in control performance is then demonstrated by using a population of varying numbers of neurons for similar control in the Nengo Neural Engineering Framework, on CPU and on Intel’s Loihi neuromorphic chip. Finally, linear control is demonstrated for four multi-linked pendula on cart systems, using a population of neurons in Nengo, followed by an implementation of the same on Loihi. This study compares LQR control in the NEF using 7 control and 7 neuromorphic performance metrics, followed by a comparison with other conventional spiking and non-spiking controllers.

In neuroscience, neural engineering, and biomedical engineering, electroencephalography (EEG) is widely used because of its non-invasiveness, high temporal resolution, and affordability. However, noise and physiological artifacts, such as cardiac, myogenic, and ocular artifacts, frequently contaminate raw EEG data. Deep learning (DL)-based denoising techniques can reduce or eliminate these artifacts, which degrade the EEG signal. Despite these techniques, significant artifacts can still hinder the performance, making noise removal a major requirement for accurate EEG analysis. Furthermore, for strong artifact removal, an Optimized Hierarchical 1D Convolutional Neural Network (1D CNN) is introduced. For effective feature extraction, the hierarchical CNN combines max-pooling, ReLU activation, and adaptive convolutional windows. An Annealed Grasshopper Algorithm (AGA) is employed to optimize the network parameters, further improving artifact removal. To ensure comprehensive exploration and convergence toward ideal CNN settings, AGA combines the fine-tuning accuracy of Simulated Annealing (SA) with the global exploration capabilities of the Grasshopper Optimization Algorithm (GOA). By utilizing a hybrid technique, the network can more effectively eliminate artifacts from various hierarchical levels, leading to a notable improvement in signal clarity and overall accuracy. The cleaned EEG data is represented by the recovered features in the last dense layer of the Hierarchical 1D CNN, which employs a sigmoid function. Based on experimental results, the proposed method achieved a PSNR of 29.5dB, MAE of 11.32, RMSE of 0.011, and CC of 0.93, which outperforms prior works. The proposed method can improve the precision of EEG artifact removal, which is a useful addition to biomedical signal processing and neuro-engineering.

Fourier ptychographic microscopy (FPM) is a computational imaging technique that achieves high-resolution complex amplitude reconstruction across a large field of view. However, conventional FPM is fundamentally limited to regions near the optical axis due to violations of the shift-invariance assumption in off-axis areas, resulting in challenging edge-of-field-of-view reconstructions. We propose neural pupil engineering FPM, termed NePE-FPM, a physical model that dynamically shifts the pupil function position during reconstruction instead of fixing the pupil function in the center of the pupil plane. NePE-FPM engineers the pupil function using an implicit neural representation with multi-resolution hash encoding, enabling continuous, smooth shifting of the pupil function without introducing additional physical parameters. By optimizing a feature-domain loss function, NePE-FPM adaptively filters Fourier-space information from low-resolution measurements, achieving accurate off-axis reconstruction without modeling off-axis propagation. Experimental results demonstrate isotropic resolution of 1149 lp/mm across an 11mm2 FOV using a 4×/0.1NA objective. The NePE-FPM bridges the gap between theoretical FPM capabilities and practical whole-slide imaging demands.

Haptic feedback is crucial for enhancing virtual immersion, but a neural coding mechanism that correlates the vibration frequency with surface roughness in haptic substitution remains unknown, which hinders the development of tribologically driven haptic interfaces. To address this limitation, this study models cross-modal neural coupling between mechanical vibrations and roughness systematically through double-blind experiments, event-related potential analysis, and electroencephalography (EEG) space-time modeling based on the long short-term memory (LSTM) method. By dynamically extracting the spatiotemporal dependence of the EEG signals by the LSTM method and quantifying neural representation similarity using Euclidean distances, this study reveals that cortical responses activated by specific vibration frequencies are highly consistent with natural roughness perception. In addition, the results of the behavioral verification confirm neurobehavioral consistency in perceptual equivalence. The results also show that vibration-touch substitution can simulate roughness perception through frequency-tuned neural coding. Further, this study proposes a cortical response-aligned haptic framework that provides a theoretical paradigm for virtual reality and teleoperation applications, thus advancing tribological cross-modal neural engineering.

Objective. Deep learning has emerged as a powerful approach for decoding electroencephalography (EEG)-based brain-computer interface (BCI) signals. However, its effectiveness is often limited by the scarcity and variability of available training data. Existing data augmentation methods often introduce signal distortions or lack physiological validity. This study proposes a novel augmentation strategy designed to improve generalization while preserving the underlying neurophysiological structure of EEG signals.Approach. We propose Background EEG Transform (BGTransform), a principled data augmentation framework that leverages the neurophysiological dissociation between task-related activity and ongoing background EEG. In contrast to existing methods, BGTransform generates new trials by selectively perturbing the background EEG component while preserving the task-related signal, thus enabling controlled variability without compromising class-discriminative features. We applied BGTransform to three publicly available EEG-BCI datasets spanning steady-state visual evoked potential and P300 paradigms. The effectiveness of BGTransform is evaluated using several widely adopted neural decoding models under three training regimes: (1) without augmentation (baseline model), (2) with conventional augmentation methods, and (3) with BGTransform.Main results. Across all datasets and model architectures, BGTransform consistently outperformed both baseline models and conventional augmentation techniques. Compared to models trained without BGTransform, it achieved average classification accuracy improvements of 2.45%-15.52%, 4.36%-17.15% and 7.55%-10.47% across the three datasets, respectively. In addition, BGTransform demonstrated greater robustness across subjects and tasks, maintaining stable performance under varying recording conditions.Significance. BGTransform provides a principled and effective approach to augmenting EEG data, informed by neurophysiological insight. By preserving task-related components and introducing controlled variability, the method addresses the challenge of data sparsity in EEG-BCI training. These findings support the utility of BGTransform for improving the accuracy, robustness, and generalizability of deep learning models in neural engineering applications.

An interactive bidirectional relationship between periodontitis and diabetes poses great challenges for the treatment of diabetic periodontitis in clinical practice. The hyperglycemic inflammatory periodontal microenvironment is characterized by oxidative damage, chronic invasive infection, excessive inflammation, unbalanced immunomodulation, progressive neuropathy, diabetic vasculopathy, and uncoupled bone resorption and formation responses. The neuromodulation strategy holds great potential to mediate and coordinate temporally the complex microenvironment for diabetic periodontal regeneration. Herein, natural grape seed polyphenols (GSPs) were used to modify the surface of exosomes derived from Schwann cells (SC Exo) through polyphenolic polymerization and noncovalent interactions. The GSP-coated SC Exo (GSP@SC Exo) system demonstrated antibacterial and antioxidative properties. Furthermore, these GSP@SC Exo nanoparticles could induce axonal growth of dorsal root ganglia explants, production of anti-inflammatory factors, M2 polarization of macrophages, tube formation of human umbilical vein endothelial cells, and human periodontal ligament stem cell osteogenic differentiation in vitro. Additionally, diabetic periodontal destruction was reversed through nerve regeneration, anti-inflammatory effect, immunomodulation, neovascularization, and alveolar bone formation in vivo. Therefore, this study provided a facile and effective neural engineering strategy for guiding the treatment of diabetic periodontitis in clinical applications.

Abstract It is with great pleasure that we present the academic proceedings of the First International Conference on Cyborg and Bionic Systems, hosted by the journal Cyborg and Bionic Systems. This symposium will be held from July 24 to 26, 2025, in Singapore, aiming to establish a free, open, and inclusive premier platform for global scholars. Bringing together experts, scholars, students, and industry pioneers in robotics, biomedical engineering, neural engineering, and related fields, the symposium will catalyze breakthrough advancements and practical applications in Cyborg and Bionic Systems. Against the backdrop of rapidly evolving bionic technologies, ICCBS 2025 serves as a vital platform for interdisciplinary dialogue and collaborative innovation. The symposium focuses on emerging trends and cutting-edge developments in bionic robotics, wearable robotics, neural engineering, medical imaging, biomaterials, biosensors, organ-on-a-chip, and bio-self-assembly/tissue engineering. Esteemed scientists who have made significant contributions to life-like technologies have been invited to deliver keynote speeches, offering profound insights into the future convergence of robotics, biomedical engineering, neural engineering, and real-world applications. This event is not only a global stage for showcasing pioneering achievements in life-like systems but also an innovative hub fostering international academic collaboration. List of Chair, Co-Chair, Programme Chair and Programme CO-Chair are available in this PDF.

Photovoltaic fault detection algorithm using ensemble learning enhanced with deep neural network feature engineering

Hearing impairment not only affects language communication and social integration, but is also closely related to a variety of neuropsychiatric problems such as cognitive decline, increased risk of depression, and increased incidence of Alzheimers disease. With the rapid development of neural engineering and sensory substitution technology, combining optogenetics and brain-computer interface (BCI) provides a new technical path for high-fidelity auditory reconstruction. This paper reviews the current application status of optogenetics in the auditory system, explores the key biological mechanisms, representative research results, and system implementation solutions, and analyzes the main challenges and potential solutions this technology faces. Multiple animal model experiments verified that optical stimulation is superior to traditional electrical cochlear stimulation in frequency resolution, spatial positioning, and temporal accuracy, showing the possibility of becoming the next generation of artificial hearing systems. Finally, combined with current progress, this paper points out that the optogenetic auditory brain-computer interface system is expected to achieve multi-channel closed-loop control and promote its clinical application in high-precision cochlear implants and personalized neural prostheses. This line of research offers valuable conceptual grounding for developing next-generation auditory neuroprostheses based on precise neural modulation.

Steady-state visual evoked potential (SSVEP), as a non-invasive EEG signal, is widely used in brain-computer interfaces (BCI) due to its high signal-to-noise ratio and high temporal resolution. With advancements in neural engineering, SSVEP-BCI technology has become one of the important research directions in the fields of medicine, rehabilitation and human-computer interaction. Despite its potential, SSVEP-BCI systems still face challenges such as signal stability, individual variability, system portability and interaction naturalness. This paper explores the relationship between SSVEP and brain function, analyzes its role in cognitive tasks, and evaluates the mechanism, current status, and challenges of SSVEP-BCI systems across various fields. By reviewing relevant literature, it examines the mechanism of SSVEP generation, its application in cognitive neuroscience, and the integration of SSVEP-BCI with other EEG signals (e.g., P300, motor imagery (MI), and electrooculogram (EOG)). Studies have indicated that the SSVEP-BCI system has high application value in brain-computer interaction, especially in the medical field, where it has been widely used in the rehabilitation of patients with movement disorders and the control of assistive devices. However, current studies face challenges such as decoding accuracy, system stability, and individual variability. By focusing on multimodal integration, deep learning, and wearables, future SSVEP-BCI developments aim to enhance accuracy, stability, and user experience, broadening their applications in medicine, industry, and entertainment.

Objective.While brain-computer interfaces (BCIs) can restore basic communication to people lacking muscle control, they cannot yet restore actions that require the extremely high reliability of natural (i.e. muscle-based) actions. Most BCI research focuses on neural engineering; it seeks to improve the measurement and analysis of brain signals. But neural engineering alone cannot make BCIs reliable.Approach.A BCI does not simply decode brain activity; it enables its user to acquire a skill that is produced not by nerves and muscles but rather by the BCI. Thus, BCI research should focus also on neuroscience; it should seek to develop BCI skills that emulate natural skills.Main results.A natural skill is produced by a network of neurons and synapses that may extend from cortex to spinal cord. This network has been given the nameheksor, from the ancient Greek wordhexis. A heksor changes through life; it modifies itself as needed to maintain the key features of its skill, the attributes that make the skill satisfactory. Heksors overlap; they share neurons and synapses. Through their concurrent changes, heksors keep neuronal and synaptic properties in anegotiated equilibriumthat enables each to produce its skill satisfactorily. A BCI-based skill is produced by asynthetic heksor, a network of neurons, synapses, and software that produces a BCI-based skill and should change as needed to maintain the skill's key features.Significance.A synthetic heksor shares neurons and synapses with natural heksors. Like natural heksors, it can benefit from multimodal sensory feedback, using signals from multiple brain areas, and maintaining the skill's key features rather than all its details. A synthetic heksor also needs successful co-adaptation between its central nervous system and software components and successful integration into the negotiated equilibrium that heksors establish and maintain. With due attention to both neural engineering and neuroscience, BCIs could become as reliable as muscles.

Autonomous vehicles are expected to reduce traffic accident casualties, as driver distraction accounts for 90% of accidents. These vehicles rely on sensors and controllers to operate independently, requiring robust security mechanisms to prevent malicious takeovers. This research proposes a novel approach to assessing the impact of cyber-attacks on autonomous vehicles and their surroundings, with a strong focus on prioritizing human safety. The system evaluates the severity of incidents caused by attacks, distinguishing between different events—for example, a pedestrian injury is classified as more critical than a collision with an inanimate object. By integrating deep neural network technology with feature engineering, the proposed system provides a comprehensive impact assessment. It is validated using metrics such as MAE, loss function, and Spearman’s correlation through experiments on a dataset of 5410 samples. Beyond enhancing autonomous vehicle security, this research contributes to real-world attack impact assessment, ensuring human safety remains a priority in the evolving autonomous landscape.

Electroencephalography (EEG) plays a crucial role in brain-computer interfaces (BCIs) and neurological diagnostics, but its real-world deployment faces challenges due to noise artifacts, missing data, and high annotation costs. We introduce SSL-SE-EEG, a framework that integrates Self-Supervised Learning (SSL) with Squeeze-and-Excitation Networks (SE-Nets) to enhance feature extraction, improve noise robustness, and reduce reliance on labeled data. Unlike conventional EEG processing techniques, SSL-SE-EEG transforms EEG signals into structured 2D image representations, suitable for deep learning. Experimental validation on MindBigData, TUH-AB, SEED-IV and BCI-IV datasets demonstrates state-of-the-art accuracy (91% in MindBigData, 85% in TUH-AB), making it well-suited for real-time BCI applications. By enabling low-power, scalable EEG processing, SSL-SE-EEG presents a promising solution for biomedical signal analysis, neural engineering, and next-generation BCIs. The code is available at https://github.com/roycmeghna/SSL_SE_EEG_EMBC25.

Poly(vinylidene fluoride) (PVDF)-based piezoelectric materials have emerged as a transformative platform in tissue engineering due to their unique ability to mimic endogenous bioelectric signals, which play pivotal roles in cellular behaviors, such as proliferation, differentiation, and tissue regeneration. This review comprehensively explores the structural polymorphism, processing techniques, and electromechanical properties of PVDF and its copolymers, emphasizing their superior piezoelectric coefficients, biocompatibility, and adaptability to diverse fabrication methods. The intrinsic piezoelectricity of PVDF, driven by its polar β-phase, enables dynamic responses to mechanical stimuli, such as physiological movements or external forces, generating localized electrical potentials that modulate critical signaling pathways to enhance tissue repair. Applications span multiple organs: in bone regeneration, PVDF scaffolds promote osteogenesis through mechanoelectrical coupling; in neural engineering, they facilitate axonal growth and myelination; in cardiac repair, they synchronize cardiomyocyte contraction; and in skin healing, they accelerate re-epithelialization and angiogenesis. Despite these advances, challenges persist, including optimizing piezoelectric output, ensuring long-term biocompatibility, and achieving controlled biodegradability. Future directions highlight the integration of PVDF with smart functionalities and the exploration of organ-specific signaling mechanisms to advance clinical translation. This work underscores the potential of PVDF-based materials as multifunctional platforms for next-generation regenerative therapies.

Brain science research is considered the crown jewel of 21st-century scientific research; the United States, the United Kingdom, and Japan have elevated brain science research to a national strategic level. This study employs bibliometric analysis and knowledge graph visualization to map global trends, research hotspots, and collaborative networks in brain science, providing insights into the field's evolving landscape and future directions. We analyzed 13,590 articles (1990-2023) from the Web of Science Core Collection using CiteSpace and VOSviewer. Metrics included publication volume, co-authorship networks, citation patterns, keyword co-occurrence, and burst detection. Analytical tools such as VOSviewer, CiteSpace, and online bibliometric platforms were employed to facilitate this investigation. The United States, China, and Germany dominated research output, with China's publications rising from sixth to second globally post-2016, driven by national initiatives like the China Brain Project. However, China exhibited limited international collaboration compared to the United States and European Union. Key journals included Human Brain Mapping and Journal of Neural Engineering, while emergent themes centered on "task analysis," "deep learning," and "brain-computer interfaces" (BCIs). Research clusters revealed three focal areas: (1) Brain Exploration (e.g., fMRI, diffusion tensor imaging), (2) Brain Protection (e.g., stroke rehabilitation, amyotrophic lateral sclerosis therapies), and (3) Brain Creation (e.g., neuromorphic computing, BCIs integrated with AR/VR). Despite China's high output, its influence lagged in highly cited scholars, reflecting a "quantity-over-quality" challenge. Brain science research is in a golden period of development. This bibliometric analysis offers the first comprehensive review, encapsulating research trends and progress in brain science. It reveals current research frontiers and crucial directions, offering a strategic roadmap for researchers and policymakers to navigate countries when planning research layouts.

The adoption of AI-driven hiring systems for modern recruitment purposes becomes widespread because these systems optimize candidate assessment and selection processes. The issues related to candidate experience together with transparency and fairness continue to represent crucial problems in such hiring approaches. The research examines methods for using neurotechnology to measure cognitive load during AI-driven recruitment processes to better understand assessment-related cognitive stress on candidates. The main purpose of this study involves investigating how cognitive load shifts when different AI-processed interview approaches are used and exploring if neurological data collections help measure candidate involvement and anxiety levels effectively. Users participated in an experimental data collection phase followed by statistical analysis to accomplish the research findings. Participating candidates went through automated interviews with structured and unstructured questioning while brain signals were tracked through electroencephalography (EEG) and eye-tracking and heart rate variability (HRV) monitored their cognitive load. The researchers added self-report surveys to the study as supplemental data collection after participants finished interviewing with AI. The research data shows that AI interview systems which operate using only computer-generated questions produce increased cognitive challenges than traditional human-assisted virtual screening. Job candidates developed elevated cognitive stress when working with AI systems which failed to show their decision-making parameters. Candidate performance improved when AI systems gave live feedback along with descriptions of evaluation metrics during the interview process. An evaluation of neurotechnological data during AI-driven hiring systems provides practical information for optimizing both candidate interviews and recruitment assessment frameworks. Research demonstrates how neurotechnologies can boost the development of AI-based hiring systems which become more favorable for candidates while remaining transparent in their evaluations. Researchers should investigate ethical factors with potential biases alongside scalability prospects in AI recruitment procedures to maintain ethical and fair hiring processes.

A century and a half of neuroscience has yielded many divergent theories of the neurobiology of language. Two factors that likely contribute to this situation include (a) conceptual disagreement about language and its component processes, and (b) intrinsic inter-individual variability in the topography of language areas. Recent functional magnetic imaging (fMRI) studies of small numbers of intensively scanned individuals have argued that a language-selective brain network emerges from correlations (individualized functional connectomics, iFC) in task-free (e.g., rest) or task-regressed activation timecourses. Here we test this hypothesis at scale and evaluate its practical utility for task-agnostic language localization: we apply iFC separately to each of 1,971 (fMRI) scanning sessions (1,199 unique brains), each consisting of diverse tasks. We find that iFC indeed reveals a left-lateralized frontotemporal network that is more stable within individuals than between them, robust to the granularity of the parcellation, and selective for language. These results support the hypothesis that this network is a key structure in the functional organization of the adult brain and show that it can be recovered retrospectively from arbitrary imaging data, with implications for neuroscience, neurosurgery, and neural engineering.