Last updated on 24.05.2023. | Please note this schedule is tentative and subject to change. Please check back for updates.
Recording neuronal activity triggered by electrical impulses is a powerful tool in neuroscience research and neural engineering. It is often applied in acute electrophysiological experimental settings to record compound nerve action potentials. However, the elicited neural response is often distorted by electrical stimulus artifacts, complicating subsequent analysis. In this talk, we present a model to better understand the effect of the selected recording amplifier configuration and the location of the ground electrode in a practical electrophysiological nerve setup. We present simulations results that show that the stimulus artifact can be reduced by more than an order of magnitude if the placement of the ground electrode, its impedance, and the amplifier configuration is optimized. We also present experimentally results which demonstrate the effects in three different settings, in-vivo and in-vitro.
Recent advances in Quantum Technologies have led to unheard-of levels of sensitivity and precision, allowing not only to peer into the farthest reaches of the universe but also giving a better understanding of our innermost selves - via a growing understanding of the bio-magnetic nature of the human body. Medical Imaging Technology, however, lags behind these developments by many decades. We seek the development of a novel quantum-driven medical imaging platform promising to offer non-invasive, portable, and affordable deep-brain imaging as well as insights into tissue health on the sub-cellular level.
Brain computer interfaces (BCI) offer high precision for treating brain disorders, but there is a lack of standards for integrating data and estimating features using machine learning. To address this issue, we present a software framework called py_neuromodulation that combines adaptive deep brain stimulation and connectomics to facilitate brain circuit discovery for brain implants. We tested this platform on data from 73 patients with movement disorders, depression, and epilepsy. We identified Network targets for emotion decoding in depression patients, performed generalized movement decoding across cohorts with Parkinson’s disease from the US, Europe, and China, and improved seizure detection in responsive neurostimulation for epilepsy. The platform's capabilities aid precision medicine by enabling personalized therapy for patients with brain disorders.
Neural probes are crucial for diagnosis and treatment, as well as in prosthetic devices. However, the durability of these probes is limited due to scarring and material deterioration, which can lead to a decrease in signal quality. While materials, coatings, flexibility, and surgical techniques are explored to overcome these challenges, a systematic and industry-compatible process for long-term testing of new probe technologies is still lacking. The development of objective quality criteria for intracortical implant testing could reduce the risk for developers and improve treatment of patients suffering from neural diseases. Therefore, we are developing quality criteria, ranging from device characterization to data analysis, to enable reliable assessment of long-term functionality and further improve intracortical longevity.
This talk will focus on the validation of in-ear EEG using full scalp EEG biomarkers. Over a series of experiments, we have used our in-ear EEG sensor to reproduce standard EEG biomarkers such as such as ASSR, SSVEO, SSEP, HEOG, Alpha etc. as well as sleep-based biomarkers such as K-complexes, sleep spindles, slow waves, arousal and REM. Our results showed that in-ear EEG compares favourably to the gold standard scalp EEG in the detection of these biomarkers. This talk will explain in detail the types of biomarkers studied and to what extent they were reproduced, as well as protocols used to elicit them and analysis methods.
Prior research indicates that resting state theta power predicts response to hypnosis treatment, suggesting the possibility that neurofeedback to enhance theta power might enhance response to clinical hypnosis. To examine this possibility in a pilot study, 22 individuals with multiple sclerosis and pain or fatigue were randomly assigned to five sessions of hypnosis treatment alone or five sessions of hypnosis combined with 10 sessions of neurofeedback. The results indicated larger improvements in pain intensity, sleep disturbance, and depression in participants who received the combined treatment than who received HYP alone. No between-group differences in fatigue were found. The results are consistent with the possibility that NF could be used to enhance response to hypnosis interventions.
We demonstrate the use of epitaxial graphene on silicon carbide on silicon as dry EEG sensors for a wearable brain-machine interface (BMI) system which is reliable upon long-term usage. The produced sensors are wafer- thin, highly biocompatible, and show a lower contact impedance as compared to their bulkier commercial sensor counterparts and are are remarkably resilient to corrosion in saline environments. The BMI evaluation used a steady -state visual evoked potential paradigm (SSVEP), presenting the individual with up to 6 command options to control hands-free the movement of a robotic quadruped. We have demonstrated up to 94% accuracy in the lab using 6 graphene sensors mounted on a rudimental helmet in the presence of 5mm length hair.
Non-invasive brain stimulation (NIBS) is widely used for neuromodulatory purposes. When applied in sleep medicine, the main hypothesis explaining its effects concerns the induction of neural plasticity and the modulation of the strength of synaptic connections between the brain areas involved in sleep disorders. Recently, a multi-database-based search converges on the evidence that NIBS is safe and effective in insomnia, restless legs syndrome, and sleep deprivation-related cognitive deficits; limited data are available for sleep bruxism and REM sleep behavior disorder, whereas no relevant effect was observed in obstructive sleep apnea syndrome and narcolepsy. Some limitations, especially regarding the population studied and the stimulation protocols adopted, need to be considered, although the development of individually tailored neuromodulatory techniques is promising.
We will present a new binocular laser pupillometer, aimed to study the pupil light reflex with unprecedented precision. The laser pupillometer provides for the first time precise spatially-selective illumination of the eye and stimulation of the retina. It combines visible and infrared laser light, the former used for stimulus, and the latter for providing pointing information. The spatial selectivity provides order-of-magnitude increase in the neural information that can be extracted from measurements of the pupil diameter dynamics, and thus can lead to novel methodologies for addressing brain function and developing new medical diagnostic tools.
This presentation will discuss the use of Electroencephalography (EEG) technology in Brain Computer Interface (BCI) applications. The focus will be on the challenges of obtaining sufficient EEG data for training deep learning models and an advanced methodology for generating synthetic data using denoising diffusion probabilistic models. The proposed approach involves generating synthetic data from emotionally labeled EEG recordings and electrode-frequency distribution maps (EFDMs). The validity of the synthetic data generated was assessed through a qualitative and quantitative comparison with real EEG data. The proposed methodology has potential implications for creating large, publicly available synthetic EEG datasets, without privacy concerns.
Sense element engagement theory provides a foundation for improving neuroprosthetic and neuroenhancement technology. Despite technological advances, the poor quality and utility of visual experience produced by stimulating visual cortex have changed little in 65 years. The theory describes prosthetic visual experience as a projection of patterns created by specified neural interactions and suggests a novel means for producing continuous visual forms. A neuromorphic device which uses population recordings to modulate the amplitude of intermittent cortical stimulation and is predicted to yield continuous visual forms has been integrated with V1 neural network simulations. Network spike distributions produced by simulations with the device are very similar to those produced by network-only simulations, and the number of phosphenes is classified very accurately using these distributions.
MindRove is dedicated to developing cutting-edge biosensing tools for the human-computer interaction field. Our innovative solutions have been utilized in various industries, including exoskeleton development, brain research, mental training, and VR. MindRove has 3 types of EEG devices and 2 types of EMG-based armband like sensors. Our EEG devices are versatile and easy to use. We have developed unique dry-electrodes. And we are in the middle of another development that we will be patented. We won a grant from ESA (European Space Agency) BIC Hungary to develop a neurofeedback application for astronaut training. We have an another serious space related project that we can announce at the NBT event.
Research of the subjective concept of sleepiness produced many controversial questions, including such questions as whether the physiological component of subjective feeling of sleepiness can be precisely defined in scientific terms, and whether the reliable physiological markers of subjective sleepiness (e.g., its EEG marker) can be identified. In the study of afternoon napping attempts of university students, we revealed the EEG-underpinnings of subjectively assessed sleepiness and concluded that the self-reports of these students on excessive daytime sleepiness can be trusted because 1) they demonstrated the signs of elevated sleep pressure during such napping attempts and 2) they are chronically sleep deprived during weekdays (when they are attending classes).
The Brain-Computer Interfaces (BCI) field is currently experiencing a momentum, attracting both researchers and hackers. At the same time, a growing number of people rely on the thriving Python datascience and machine learning ecosystem. Yet, until recently, there was no fully open source Python framework for building BCIs. Timeflux aims to fill this gap. Attendees will learn what can BCIs be used for, how they actually work, what are the core concepts driving Timeflux, how to describe processing pipelines, how to create interfaces available from a web browser, and how to easily implement their own algorithms for both offline and online use. We will conclude with a practical example: a mind-controlled virtual keyboard.
Artificial intelligence (AI) is revolutionizing the field of human-computer interfaces (HCI) by enhancing their capabilities and expanding their potential applications. AI algorithms analyze and interpret signals, gestures enabling more accurate and efficient communication between the human and external devices. But can AI change the course of Brain Computer Interfaces, especially this relying on surface EEG? By harnessing the power of AI, BCIs are becoming increasingly sophisticated, user-friendly, and capable of transforming the lives of individuals with disabilities, enhancing human cognition, and unlocking new frontiers in human-machine interaction.
Due to demographic change, dementia is a huge challenge in our society. We are exploring the potentials of Brain-Computer Interfaces (BCIs) within the METIS digital platform. METIS is a research-transfer project, which employs a multistage, risk-factor based approach to create individualized outpatient treatments for neurodegenerative diseases. For this, we designed, a hybrid electrode, allowing for non-intrusive, easy-to-perform BCI recordings, aiming to objectively assess treatment efficacy. Through AI integration and wearable technologies, we aim to fine-tune patient profiles and intervention strategies. BCIs present a novel objective measure for the success of dementia intervention strategies, offering an opportunity to positively influence disease progression.
BitBrain is a novel learning mechanism based upon a novel synthesis of ideas from sparse coding, computational neuroscience and information theory that enables fast, adaptive learning and accurate, robust inference. This method is designed to be implemented efficiently on neuromorphic devices as well as on more conventional CPU and memory architectures. The resulting inference mechanism has excellent performance on benchmarks such as MNIST and EMNIST, achieving state-of-the-art accuracy with single-pass learning. It provides a unique combination of single-pass, single-shot and continuous supervised learning; following a very simple unsupervised phase. Accurate classification inference that is very robust against imperfect inputs and noise has been demonstrated. These contributions make it uniquely well-suited for edge and IoT applications.
In this fNIRS hyperscanning demo, we will showcase the potential of using functional near infrared spectroscopy (fNIRS) to simultaneously record brain activity from two individuals engaged in an interaction. This fNIRS hyperscanning demo is designed with quick setup, comfort, and mobility in mind. The system is designed to be comfortable for the participants, with lightweight and flexible optodes that do not interfere with their movement or comfort during the interaction. Additionally, we will showcase the portability of the system, which can be easily moved from one location to another for on-site data collection. Overall, our demo aims to demonstrate the ease and practicality of fNIRS. By demonstrating real-time data acquisition and analysis, we aim to highlight the applicability of fNIRS hyperscanning in various research domains, including neuroscience, psychology, and social sciences.
The coherence of neural activity analysis is challenged by an increasing heterogeneity of measurement devices, analysis tools, and data formats. To overcome these challenges, we emphasize the benefits of open-source tools, community standards, and modular design principles. Specifically, we present the Collaborative Brain Wave Analysis Pipeline (Cobrawap) as an adaptable and reusable analysis workflow solution for studying cortical wave activity. By leveraging the open-source software ecosystem, our pipeline approach allows the construction of reproducible workflows through seamless chaining of method blocks. Its inherent adaptability supports the processing of data from various measurement techniques and simulations, and the reuse and extension for wide-ranging research applications. Therefore, this open and modular approach enhances automation, integrates diverse methods, and enables comprehensive cross-domain comparisons.
