Time-frequency filtering of MEG signals with matching pursuit

HYBRIDMINDS-summary and outlook of the 2023 international conference on the ethics and regulation of intelligent neuroprostheses

Neurotechnology and Artificial Intelligence (AI) have developed rapidly in recent years with an increasing number of applications and AI-enabled devices that are about to enter the market. While promising to substantially improve quality of life across various severe medical conditions, there are also concerns that the convergence of these technologies, e.g., in the form of intelligent neuroprostheses, may have undesirable consequences and compromise cognitive liberty, mental integrity, or mental privacy. Therefore, various international organizations, such as the Organization for Economic Cooperation and Development (OECD) or United Nations Educational, Scientific and Cultural Organization (UNESCO), have formed initiatives to tackle such questions and develop recommendations that mitigate risks while fostering innovation. In this context, a first international conference on the ethics and regulation of intelligent neuroprostheses was held in Berlin, Germany, in autumn 2023. The conference gathered leading experts in neuroscience, engineering, ethics, law, philosophy as well as representatives of industry, policy making and the media. Here, we summarize the highlights of the conference, underline the areas in which a broad consensus was found among participants, and provide an outlook on future challenges in development, deployment, and regulation of intelligent neuroprostheses.

A novel method based on matching pursuit decomposition of gait signals for Parkinson's disease, Amyotrophic lateral sclerosis and Huntington's disease detection

BACKGROUND AND OBJECTIVE

An accurate detection of neurodegenerative diseases (NDDs) definitely improves the life of patients and has attracted growing attention.

METHODS

In this paper, a general automatic method for detection of Parkinson's disease (PD), Amyotrophic lateral sclerosis (ALS) and Huntington's disease (HD) has been proposed based on the localized time-frequency information of gait signals. The new main part of the detection method is to obtain a small set of sparse coefficients for the local representation of gait signals with appropriate time and frequency resolution. For this purpose, a hybrid feature set based on sparse matching pursuit decomposition and two sets of nonlinear and linear features has been developed. Then, principal components of the proposed feature have been analyzed using a sparse coding classifier. Results The proposed approach has achieved high average accuracy rates of 93%, 94%, and 97% for PD, ALS, and HD detection, respectively.

CONCLUSIONS

The obtained results have indicated that combination of time and frequency information of the gait signals through adaptive localized window length in MP makes it more efficient in comparison with the existing time, frequency or other time-frequency gait parameters. The great potential of nonlinear sparse features for PD and HD detection and linear ones for ALS detection has also been shown.

Multichannel system based on a high sensitivity superconductive sensor for magnetoencephalography

We developed a multichannel system based on superconducting quantum interference devices (SQUIDs) for magnetoencephalography measurements. Our system consists of 163 fully-integrated SQUID magnetometers, 154 channels and 9 references, and all of the operations are performed inside a magnetically-shielded room. The system exhibits a magnetic field noise spectral density of approximatively 5 fT/Hz(1=2). The presented magnetoencephalography is the first system working in a clinical environment in Italy.

Stochastic Behavior of Phase Synchronization Index and Cross-Frequency Couplings in Epileptogenic Zones during Interictal Periods Measured with Scalp dEEG

The stochastic behavior of the phase synchronization index (SI) and cross-frequency couplings on different days during a hospital stay of three epileptic patients was studied for non-invasive localization of the epileptogenic areas from high density, 256-channel, scalp EEG (dEEG) recordings. The study was performed with short-duration (0–180 s), seizure-free, epileptiform-free, and spike-free interictal dEEG data on different days of three subjects. The seizure areas were localized with subdural recordings with an 8 × 8 macro-electrode grid array and strip electrodes. The study was performed in theta (3–7 Hz), alpha (7–12 Hz), beta (12–30 Hz), and low gamma (30–50 Hz) bands. A detrended fluctuation analysis was used to find the long range temporal correlations in the SI that reveals the stochastic behavior of the SI in a given time period. The phase synchronization was computed after taking Hilbert transform of the EEG data. Contour plots were constructed with 20 s time-frames using a montage of the layout of 256 electrode positions. It was found that the stochastic behavior of the SI was higher in epileptogenic areas and in nearby areas on different days for each subject. The low gamma band was found to be the best to localize the epileptic sites. Also, a stable higher pattern of SI emerged after 60–120 s in the epileptogenic areas. The cross-frequency couplings of SI in theta–gamma, beta–gamma, and alpha–gamma bands were decreased and spatial patterns were fragmented in epileptogenic areas. Combinations of an increase in the stochastic behavior of the SI and decrease in cross-frequency couplings are potential markers to assist in localizing epileptogenic areas. These findings suggest that it is possible to localize the epileptogenic areas non-invasively from a short-duration (∼180 s), seizure-free and spike-free interictal scalp dEEG recordings.

Multichannel matching pursuit of MEG signals for discriminative oscillation pattern detection in depression

Magnetoencephalography (MEG) power topography may be useful for obtaining discriminative brain activity patterns that can distinguish depressed patients from healthy control subjects at the individual level. However, the application is still limited due to the lack of adequate analysis strategies to remove artifacts from the MEG signals. In this study, the multichannel matching pursuit (MMP) method was designed; in this technique, a linear decomposition method that provides components by iteratively reanalysing a residual signal after removing previously found components,. Forty-four subjects, half depressed patients and half healthy subjects, were recruited for MEG scanning whilst watching a video of sad faces. MMP was implemented to manage multichannel, multi-trial MEG signals. The representative post-MMP analysis signals were utilised to calculate the power topography over the whole brain and designed as inputs for a Support Vector Machine (SVM) classifier. A statistically significant discriminative accuracy of 86% (p=0.002) after a permutation test was achieved. Comparing the system classification performance to that of the ensemble averaging method and the established Independent Component Analysis (ICA), we demonstrated the ability of MMP to represent critical MEG information and, in turn, to mark the abnormality of oscillatory activities under negative stimuli using images of sad faces.

Abnormal brain processing of pain in migraine without aura: A high-density EEG brain mapping study

In the present study we used high-density EEG brain mapping to investigate spatio-temporal aspects of brain activity in response to experimentally induced muscle pain in 17 patients with migraine without aura and 15 healthy controls. Painful electrical stimuli were applied to the trapezius muscle and somatosensory-evoked potentials were recorded with 128-channel EEG with and without concurrent induced tonic neck/shoulder muscle pain. At baseline, the calculated P300 dipole for single stimuli was localized in the cingulate cortex. In patients, but not in controls, the dipole changed position from baseline to the tonic muscle pain condition (z = 29 mm vs. z = −13 mm, P < 0.001) and from baseline to the post-tonic muscle pain condition (z = 29 mm vs. z = −9 mm, P < 0.001). This may be the first evidence that the supraspinal processing of muscle pain is abnormal in patients with migraine without aura.

Power spectral density changes and language lateralization during covert object naming tasks measured with high-density EEG recordings

Our objective was to study changes in EEG time-domain power spectral density (PSDt) and localization of language areas during covert object naming tasks in human subjects with epilepsy. EEG data for subjects with epilepsy were acquired during the covert object naming tasks using a net of 256 electrodes. The trials required each subject to provide the names of common objects presented every 4 seconds on slides. Each trial comprised the 1.0 second before and 3.0 seconds after initial object presentation. PSDt values at baseline and during tasks were calculated in the theta, alpha, beta, low gamma, and high gamma bands. The spatial contour plots reveal that PSDt values during object naming were 10-20% higher than the baseline values for different bands. Language was lateralized to left frontal or temporal areas. In all cases, the Wada test disclosed language lateralization to the left hemisphere as well.

Is the pain in chronic pancreatitis of neuropathic origin? Support from EEG studies during experimental pain

AIM: To prove the hypothesis that patients with chronic pancreatitis would show increased theta activity during painful visceral stimulation.

METHODS: Eight patients and 12 healthy controls underwent an experiment where the esophagus was electrically stimulated at the pain threshold using a nasal endoscope. The electroencephalogram (EEG) was recorded from 64 surface electrodes and “topographic matching pursuit” was used to extract the EEG information in the early brain activation after stimulation.

RESULTS: A major difference between controls and patients were seen in delta and theta bands, whereas there were only minor differences in other frequency bands. In the theta band, the patients showed higher activity than controls persisting throughout the 450 ms of analysis with synchronous brain activation between the channels. The main theta components oscillated with 4.4 Hz in the patients and 5.5 Hz in the controls. The energy in the delta (0.5-3.5 Hz) band was higher in the controls, whereas the patients only showed scattered activity in this band.

CONCLUSION: The differences in the theta band indicate that neuropathic pain mechanisms are involved in chronic pancreatitis. This has important implications for the understanding and treatment of pain in these patients, which should be directed against drugs with effects on neuropathic pain disorders.