Journal of Neuroscience Methods最新文献

Background

Electroencephalography (EEG) and electrocorticography (ECoG) recordings have been used to decode finger movements by analyzing brain activity. Traditional methods focused on single bandpass power changes for movement decoding, utilizing machine learning models requiring manual feature extraction.

New method

This study introduces a 3D convolutional neural network (3D-CNN) model to decode finger movements using ECoG data. The model employs adaptive, explainable AI (xAI) techniques to interpret the physiological relevance of brain signals. ECoG signals from epilepsy patients during awake craniotomy were processed to extract power spectral density across multiple frequency bands. These data formed a 3D matrix used to train the 3D-CNN to predict finger trajectories.

Results

The 3D-CNN model showed significant accuracy in predicting finger movements, with root-mean-square error (RMSE) values of 0.26–0.38 for single finger movements and 0.20–0.24 for combined movements. Explainable AI techniques, Grad-CAM and SHAP, identified the high gamma (HG) band as crucial for movement prediction, showing specific cortical regions involved in different finger movements. These findings highlighted the physiological significance of the HG band in motor control.

Comparison with existing methods

The 3D-CNN model outperformed traditional machine learning approaches by effectively capturing spatial and temporal patterns in ECoG data. The use of xAI techniques provided clearer insights into the model's decision-making process, unlike the "black box" nature of standard deep learning models.

Conclusions

The proposed 3D-CNN model, combined with xAI methods, enhances the decoding accuracy of finger movements from ECoG data. This approach offers a more efficient and interpretable solution for brain-computer interface (BCI) applications, emphasizing the HG band's role in motor control.

The prevalence of brain tumor disorders is currently a global issue. In general, radiography, which includes a large number of images, is an efficient method for diagnosing these life-threatening disorders. The biggest issue in this area is that it takes a radiologist a long time and is physically strenuous to look at all the images. As a result, research into developing systems based on machine learning to assist radiologists in diagnosis continues to rise daily. Convolutional neural networks (CNNs), one type of deep learning approach, have been pivotal in achieving state-of-the-art results in several medical imaging applications, including the identification of brain tumors. CNN hyperparameters are typically set manually for segmentation and classification, which might take a while and increase the chance of using suboptimal hyperparameters for both tasks. Bayesian optimization is a useful method for updating the deep CNN's optimal hyperparameters. The CNN network, however, can be considered a "black box" model because of how difficult it is to comprehend the information it stores because of its complexity. Therefore, this problem can be solved by using Explainable Artificial Intelligence (XAI) tools, which provide doctors with a realistic explanation of CNN's assessments. Implementation of deep learning-based systems in real-time diagnosis is still rare. One of the causes could be that these methods don't quantify the Uncertainty in the predictions, which could undermine trust in the AI-based diagnosis of diseases. To be used in real-time medical diagnosis, CNN-based models must be realistic and appealing, and uncertainty needs to be evaluated. So, a novel three-phase strategy is proposed for segmenting and classifying brain tumors. Segmentation of brain tumors using the DeeplabV3+ model is first performed with tuning of hyperparameters using Bayesian optimization. For classification, features from state-of-the-art deep learning models Darknet53 and mobilenetv2 are extracted and fed to SVM for classification, and hyperparameters of SVM are also optimized using a Bayesian approach. The second step is to understand whatever portion of the images CNN uses for feature extraction using XAI algorithms. Using confusion entropy, the Uncertainty of the Bayesian optimized classifier is finally quantified. Based on a Bayesian-optimized deep learning framework, the experimental findings demonstrate that the proposed method outperforms earlier techniques, achieving a 97 % classification accuracy and a 0.98 global accuracy.

Background

Transcranial magnetic stimulation (TMS) is a valuable technique for assessing the function of the motor cortex and cortico-muscular pathways. TMS activates the motoneurons in the cortex, which after transmission along cortico-muscular pathways can be measured as motor-evoked potentials (MEPs). The position and orientation of the TMS coil and the intensity used to deliver a TMS pulse are considered central TMS setup parameters influencing the presence/absence of MEPs.

New method

We sought to predict the presence of MEPs from TMS setup parameters using machine learning. We trained different machine learners using either within-subject or between-subject designs.

Results

We obtained prediction accuracies of on average 77 % and 65 % with maxima up to up to 90 % and 72 % within and between subjects, respectively. Across the board, a bagging ensemble appeared to be the most suitable approach to predict the presence of MEPs.

Conclusions

Although within a subject the prediction of MEPs via TMS setup parameter-based machine learning might be feasible, the limited accuracy between subjects suggests that the transfer of this approach to experimental or clinical research comes with significant challenges.

Preterm infants are a high-risk group for brain injury, and it is important to evaluate the neurological recovery of preterm infants. Therefore, this paper evaluates the neurological recovery in preterm infants at high risk of brain injury by amplitude-integrated EEG and GMs scale. The study collected basic information on preterm infants and performed amplitude integrated EEG examination and GMs scale evaluation. Amplitude integrated EEG examination attaches electrodes using multielectrode arrays onto specific areas of the premature head to record brain wave activity to monitor electrical activity in the preterm brain in real time and amplify and process through the signals received by the electrodes to obtain more detailed EEG data. The GMs scale evaluates the developmental and functional status of the child and allows an objective assessment of the development and recovery of neurological function by observing their performance in motor, language, cognition, and social interaction. Analysis of the data by statistical processing. The results showed that early brain injury was evident in high-risk infants. Amplitude integrated EEG parameters can have some predictive value for brain injury. There were also differences in GMs scale assessment between brain injury and non-brain injury. Amplitude integrated EEG combined with GMs scale has certain value in predicting brain injury and can provide an important basis for early intervention in children with preterm brain injury and help to improve their neurodevelopmental outcome.

Background

Chronobiology is the scientific field focused on studying periodicity in biological processes. In mammals, most physiological variables exhibit circadian rhythmicity, such as metabolism, body temperature, locomotor activity, and sleep. The biological rhythmicity can be statistically evaluated by examining the time series and extracting parameters that correlate to the period of oscillation, its amplitude, phase displacement, and overall variability.

New method

We have developed a library called CircadiPy, which encapsulates methods for chronobiological analysis and data inspection, serving as an open-access toolkit for the analysis and interpretation of chronobiological data. The package was designed to be flexible, comprehensive and scalable in order to assist research dealing with processes affected or influenced by rhythmicity.

Results

The results demonstrate the toolkit's capability to guide users in analyzing chronobiological data collected from various recording sources, while also providing precise parameters related to the circadian rhythmicity.

Comparison with existing methods

The analysis methodology from this proposed library offers an opportunity to inspect and obtain chronobiological parameters in a straightforward and cost-free manner, in contrast to commercial tools.

Conclusions

Moreover, being an open-source tool, it empowers the community with the opportunity to contribute with new functions, analysis methods, and graphical visualizations given the simplified computational method of time series data analysis using an easy and comprehensive pipeline within a single Python object.

Background

Zebrafish are vertebrates with a high potential of regeneration after injury in the central nervous system. Therefore, they have emerged as a useful model system for studying traumatic spinal cord injuries.

New Method

Using larval zebrafish, we have developed a robust paradigm to model the effects of anterior spinal cord injury, which correspond to the debilitating injuries of the cervical and thoracic regions in humans. Our new paradigm consists of a more anterior injury location compared to previous studies, a modified behavioral assessment using the visual motor response, and a new data analysis code.

Results

Our approach enables a spinal cord injury closer to the hindbrain with more functional impact compared to previous studies using a more posterior injury location. Results reported in this work reveal recovery over seven days following spinal cord injury.

Comparing with existing Methods

The present work describes a modified paradigm for the in vivo study of spinal cord regeneration after injury using larval zebrafish, including an anterior injury location, a robust behavioral assessment, and a new data analysis software.

Conclusions

Our findings lay the foundation for applying this paradigm to study the effects of drugs, nutrition, and other treatments to improve the regeneration process.

Background

The conventional “whole-cell patch-clamp” recording technique is widely used to measure the resting membrane potential (V_M) and to dissect the underlying membrane ionic conductances in isolated vascular endothelial cells.

New method

Herein, we assessed whether the automated patch-clamp (APC) technology, which replaces the traditional patch-pipette with a planar substrate to permit researchers lacking formal training in electrophysiology to generate large amounts of data in a relatively short time, can be used to characterize the bioelectrical activity of vascular endothelial cells. We assessed whether the Port-a-Patch planar patch-clamp system, which is regarded as the smallest electrophysiological rig available on the market, can be used to measure the V_M and resting membrane currents in the human cerebrovascular endothelial cell line, hCMEC/D3.

Comparison with existing methods

We demonstrated that the Port-a-Patch planar patch-clamp system provides the same values of the resting V_M as those provided by the conventional patch-clamp technique. Furthermore, the APC technology provides preliminary data demonstrating that the resting V_M of hCMEC/D3 cells is primarily contributed by Cl^- and Na⁺, as demonstrated with the patch-clamp technique for many other endothelial cell types.

Conclusions

The Port-a-Patch planar patch-clamp system can be successfully used to measure the resting V_M and the underlying membrane ionic conductances in hCMEC/D3 cells. We envisage that this easy-to-use APC system could also be extremely useful for the investigation of the membrane currents that can be activated by chemical, thermal, optical, and mechanical stimuli in this cell line as well as in other types of isolated vascular endothelial cells.

Background

Assessment of taste and somatosensory perception in clinical practice lacks fast tests that are validated and reliable. Recently, a 12-item identification test for taste and oral trigeminal perception, and its shorter version, the Seven-iTT, was developed. The objectives of this study were to evaluate its test-retest reliability and establish normative data.

New method

Two-hundred participants (120 women, 80 men) with a good sense of taste performed a whole-mouth identification test using 12 filter-paper strips impregnated with low and high concentrations of sweet, sour, salty, bitter, astringency, and spiciness. Fifty of them repeated the task, with a median interval of 122 days from the first visit. Test-retest reliability was determined using Spearman correlation and the Bland–Altman plot method.

Results

There was a significant correlation in identification score between the first and the second session for both versions of the test (r ≥ 0.28; p ≤ 0.048). The Bland–Altman plot reflected a good congruence between the results of the two sessions. Additionally, frequencies of correct identification were consistent between sessions, with women outperforming men (p = 0.005). Hypogeusia was established at Seven-iTT score of 3 of less.

Comparison with existing methods

The identification test combines taste and somatosensory perception, thus creating a more detailed diagnosis tool. Scores were correlated with self-rated taste perception.

Conclusion

The present results confirmed the applicability of Seven-iTT for a reliable, fast evaluation of taste and somatosensory perception in the general population, that can be extended to clinical practice.

Background

Rehabilitation training based on the brain-computer interface of motor imagery (MI-BCI) can help restore the connection between the brain and movement. However, the performance of most popular MI-BCI system is coarse-level, which means that they are good at guiding the rehabilitation exercises of different parts of the body, but not for the individual component.

New methods

In this paper, we designed a fine-level MI-BCI system for unilateral upper limb rehabilitation assistance. Besides, due to the low discrimination of different sample classes in a single part, a classification algorithm called spatial-temporal filtering convolutional network (STFCN) was proposed that used spatial filtering and deep learning.

Comparison with existing methods

Our STFCN outperforms popular methods in recent years using BCI IV 2a and 2b data sets.

Results

To verify the effectiveness of our system, we recruited 6 volunteers and collected their data for a four-classification online experiments, resulting in an average accuracy of 62.7 %.

Conclusion

This fine-level MI-BCI system has good appli-cation prospects, and inspires more exploration of rehabilitation in a single part of the human body.

Background

In electroencephalographic (EEG) or electrocorticographic (ECoG) experiments, visual cues are commonly used for timing synchronization but may inadvertently induce neural activity and cognitive processing, posing challenges when decoding self-initiated tasks.

New method

To address this concern, we introduced four new visual cues (Fade, Rotation, Reference, and Star) and investigated their impact on brain signals. Our objective was to identify a cue that minimizes its influence on brain activity, facilitating cue-effect free classifier training for asynchronous applications, particularly aiding individuals with severe paralysis.

Results

22 able-bodied, right-handed participants aged 18–30 performed hand movements upon presentation of the visual cues. Analysis of time-variability between movement onset and cue-aligned data, grand average MRCP, and classification outcomes revealed significant differences among cues. Rotation and Reference cue exhibited favorable results in minimizing temporal variability, maintaining MRCP patterns, and achieving comparable accuracy to self-paced signals in classification.

Comparison with existing methods

Our study contrasts with traditional cue-based paradigms by introducing novel visual cues designed to mitigate unintended neural activity. We demonstrate the effectiveness of Rotation and Reference cue in eliciting consistent and accurate MRCPs during motor tasks, surpassing previous methods in achieving precise timing and high discriminability for classifier training.

Conclusions

Precision in cue timing is crucial for training classifiers, where both Rotation and Reference cue demonstrate minimal variability and high discriminability, highlighting their potential for accurate classifications in online scenarios. These findings offer promising avenues for refining brain-computer interface systems, particularly for individuals with motor impairments, by enabling more reliable and intuitive control mechanisms.