Augmenting interictal mapping with neurovascular coupling biomarkers by structured factorization of epileptic EEG and fMRI data.

EEG-correlated fMRI analysis is widely used to detect regional BOLD fluctuations that are synchronized to interictal epileptic discharges, which can provide evidence for localizing the ictal onset zone. However, the typical, asymmetrical and mass-univariate approach cannot capture the inherent, higher order structure in the EEG data, nor multivariate relations in the fMRI data, and it is nontrivial to accurately handle varying neurovascular coupling over patients and brain regions. We aim to overcome these drawbacks in a data-driven manner by means of a novel structured matrix-tensor factorization: the single-subject EEG data (represented as a third-order spectrogram tensor) and fMRI data (represented as a spatiotemporal BOLD signal matrix) are jointly decomposed into a superposition of several sources, characterized by space-time-frequency profiles. In the shared temporal mode, Toeplitz-structured factors account for a spatially specific, neurovascular 'bridge' between the EEG and fMRI temporal fluctuations, capturing the hemodynamic response's variability over brain regions. By analyzing interictal data from twelve patients, we show that the extracted source signatures provide a sensitive localization of the ictal onset zone (10/12). Moreover, complementary parts of the IOZ can be uncovered by inspecting those regions with the most deviant neurovascular coupling, as quantified by two entropy-like metrics of the hemodynamic response function waveforms (9/12). Hence, this multivariate, multimodal factorization provides two useful sets of EEG-fMRI biomarkers, which can assist the presurgical evaluation of epilepsy. We make all code required to perform the computations available at https://github.com/svaneynd/structured-cmtf.

A deep learning model for brain segmentation across pediatric and adult populations.

Automated quantification of brain tissues on MR images has greatly contributed to the diagnosis and follow-up of neurological pathologies across various life stages. However, existing solutions are specifically designed for certain age ranges, limiting their applicability in monitoring brain development from infancy to late adulthood. This retrospective study aims to develop and validate a brain segmentation model across pediatric and adult populations. First, we trained a deep learning model to segment tissues and brain structures using T1-weighted MR images from 390 patients (age range: 2-81 years) across four different datasets. Subsequently, the model was validated on a cohort of 280 patients from six distinct test datasets (age range: 4-90 years). In the initial experiment, the proposed deep learning-based pipeline, icobrain-dl, demonstrated segmentation accuracy comparable to both pediatric and adult-specific models across diverse age groups. Subsequently, we evaluated intra- and inter-scanner variability in measurements of various tissues and structures in both pediatric and adult populations computed by icobrain-dl. Results demonstrated significantly higher reproducibility compared to similar brain quantification tools, including childmetrix, FastSurfer, and the medical device icobrain v5.9 (p-value< 0.01). Finally, we explored the potential clinical applications of icobrain-dl measurements in diagnosing pediatric patients with Cerebral Visual Impairment and adult patients with Alzheimer's Disease.

Artificial Intelligence Assistive Software Tool for Automated Detection and Quantification of Amyloid-Related Imaging Abnormalities.

Importance: Amyloid-related imaging abnormalities (ARIA) are brain magnetic resonance imaging (MRI) findings associated with the use of amyloid-ß-directed monoclonal antibody therapies in Alzheimer disease (AD). ARIA monitoring is important to inform treatment dosing decisions and might be improved through assistive software. Objective: To assess the clinical performance of an artificial intelligence (AI)-based software tool for assisting radiological interpretation of brain MRI scans in patients monitored for ARIA. Design, Setting, and Participants: This diagnostic study used a multiple-reader multiple-case design to evaluate the diagnostic performance of radiologists assisted by the software vs unassisted. The study enrolled 16 US Board of Radiology-certified radiologists to perform radiological reading with (assisted) and without the software (unassisted). The study encompassed 199 retrospective cases, where each case consisted of a predosing baseline and a postdosing follow-up MRI of patients from aducanumab clinical trials PRIME, EMERGE, and ENGAGE. Statistical analysis was performed from April to July 2023. Exposures: Use of icobrain aria, an AI-based assistive software for ARIA detection and quantification. Main Outcomes and Measures: Coprimary end points were the difference in diagnostic accuracy between assisted and unassisted detection of ARIA-E (edema and/or sulcal effusion) and ARIA-H (microhemorrhage and/or superficial siderosis) independently, assessed with the area under the receiver operating characteristic curve (AUC). Results: Among the 199 participants included in this study of radiological reading performance, mean (SD) age was 70.4 (7.2) years; 105 (52.8%) were female; 23 (11.6%) were Asian, 1 (0.5%) was Black, 157 (78.9%) were White, and 18 (9.0%) were other or unreported race and ethnicity. Among the 16 radiological readers included, 2 were specialized neuroradiologists (12.5%), 11 were male individuals (68.8%), 7 were individuals working in academic hospitals (43.8%), and they had a mean (SD) of 9.5 (5.1) years of experience. Radiologists assisted by the software were significantly superior in detecting ARIA than unassisted radiologists, with a mean assisted AUC of 0.87 (95% CI, 0.84-0.91) for ARIA-E detection (AUC improvement of 0.05 [95% CI, 0.02-0.08]; P = .001]) and 0.83 (95% CI, 0.78-0.87) for ARIA-H detection (AUC improvement of 0.04 [95% CI, 0.02-0.07]; P = .001). Sensitivity was significantly higher in assisted reading compared with unassisted reading (87% vs 71% for ARIA-E detection; 79% vs 69% for ARIA-H detection), while specificity remained above 80% for the detection of both ARIA types. Conclusions and Relevance: This diagnostic study found that radiological reading performance for ARIA detection and diagnosis was significantly better when using the AI-based assistive software. Hence, the software has the potential to be a clinically important tool to improve safety monitoring and management of patients with AD treated with amyloid-ß-directed monoclonal antibody therapies.

AI-Based Chest CT Analysis for Rapid COVID-19 Diagnosis and Prognosis: A Practical Tool to Flag High-Risk Patients and Lower Healthcare Costs.

Early diagnosis of COVID-19 is required to provide the best treatment to our patients, to prevent the epidemic from spreading in the community, and to reduce costs associated with the aggravation of the disease. We developed a decision tree model to evaluate the impact of using an artificial intelligence-based chest computed tomography (CT) analysis software (icolung, icometrix) to analyze CT scans for the detection and prognosis of COVID-19 cases. The model compared routine practice where patients receiving a chest CT scan were not screened for COVID-19, with a scenario where icolung was introduced to enable COVID-19 diagnosis. The primary outcome was to evaluate the impact of icolung on the transmission of COVID-19 infection, and the secondary outcome was the in-hospital length of stay. Using EUR 20000 as a willingness-to-pay threshold, icolung is cost-effective in reducing the risk of transmission, with a low prevalence of COVID-19 infections. Concerning the hospitalization cost, icolung is cost-effective at a higher value of COVID-19 prevalence and risk of hospitalization. This model provides a framework for the evaluation of AI-based tools for the early detection of COVID-19 cases. It allows for making decisions regarding their implementation in routine practice, considering both costs and effects.

icobrain ms 5.1: Combining unsupervised and supervised approaches for improving the detection of multiple sclerosis lesions.

Multiple sclerosis (MS) is a chronic autoimmune, inflammatory neurological disease of the central nervous system. Its diagnosis nowadays commonly includes performing an MRI scan, as it is the most sensitive imaging test for MS. MS plaques are commonly identified from fluid-attenuated inversion recovery (FLAIR) images as hyperintense regions that are highly varying in terms of their shapes, sizes and locations, and are routinely classified in accordance to the McDonald criteria. Recent years have seen an increase in works that aimed at development of various semi-automatic and automatic methods for detection, segmentation and classification of MS plaques. In this paper, we present an automatic combined method, based on two pipelines: a traditional unsupervised machine learning technique and a deep-learning attention-gate 3D U-net network. The deep-learning network is specifically trained to address the weaker points of the traditional approach, namely difficulties in segmenting infratentorial and juxtacortical plaques in real-world clinical MRIs. It was trained and validated on a multi-center multi-scanner dataset that contains 159 cases, each with T1 weighted (T1w) and FLAIR images, as well as manual delineations of the MS plaques, segmented and validated by a panel of raters. The detection rate was quantified using lesion-wise Dice score. A simple label fusion is implemented to combine the output segmentations of the two pipelines. This combined method improves the detection of infratentorial and juxtacortical lesions by 14% and 31% respectively, in comparison to the unsupervised machine learning pipeline that was used as a performance assessment baseline.

Semi-automated EEG Enhancement Improves Localization of Ictal Onset Zone With EEG-Correlated fMRI.

Objective: To improve the accuracy of detecting the ictal onset zone, we propose to enhance the epilepsy-related activity present in the EEG signals, before mapping their BOLD correlates through EEG-correlated fMRI analysis. Methods: Based solely on a segmentation of interictal epileptic discharges (IEDs) on the EEG, we train multi-channel Wiener filters (MWF) which enhance IED-like waveforms, and suppress background activity and noisy influences. Subsequently, we use EEG-correlated fMRI to find the brain regions in which the BOLD signal fluctuation corresponds to the filtered signals' time-varying power (after convolving with the hemodynamic response function), and validate the identified regions by quantitatively comparing them to ground-truth maps of the (resected or hypothesized) ictal onset zone. We validate the performance of this novel predictor vs. that of commonly used unitary or power-weighted predictors and a recently introduced connectivity-based metric, on a cohort of 12 patients with refractory epilepsy. Results: The novel predictor, derived from the filtered EEG signals, allowed the detection of the ictal onset zone in a larger percentage of epileptic patients (92% vs. at most 83% for the other predictors), and with higher statistical significance, compared to existing predictors. At the same time, the new method maintains maximal specificity by not producing false positive activations in healthy controls. Significance: The findings of this study advocate for the use of the MWF to maximize the signal-to-noise ratio of IED-like events in the interictal EEG, and subsequently use time-varying power as a sensitive predictor of the BOLD signal, to localize the ictal onset zone.

Unsupervised respiratory signal extraction from ungated cardiac magnetic resonance imaging at rest and during exercise.

We propose and evaluate a method to estimate a respiratory signal from ungated cardiac magnetic resonance (CMR) images. Ungated CMR images were acquired in five subjects who performed exercise at different intensity levels under different physiological conditions while breathing freely. The respiratory motion was estimated by applying principal components analysis (PCA). A sign correction procedure was developed to correctly define inspiration and expiration, based on either tracking of the diaphragmatic motion or estimation of the lung volume or a combination of both. Evaluation was done using a plethysmograph signal as reference. There was a good correspondence between the plethysmograph and the estimated respiratory signals. Respiratory motion was effectively captured by one of the PCA components in 88% of the cases. Moreover, the proposed method successfully estimated the respiratory phase in 91% of the evaluated slices. The pipeline is robust, admitting a slight decline in performance with increased exercise intensity. Respiratory motion was accurately estimated by means of PCA and the application of a sign correction procedure. Our method showed promising results even for acquisitions during exercise where excessive body motion occurs. The proposed method provides a way to extract the respiratory signal from ungated CMR images, at rest as well as during exercise, in a fully unsupervised fashion, which may reduce the clinician's workload drastically.

EEG-Informed Attended Speaker Extraction From Recorded Speech Mixtures With Application in Neuro-Steered Hearing Prostheses.

OBJECTIVE: We aim to extract and denoise the attended speaker in a noisy two-speaker acoustic scenario, relying on microphone array recordings from a binaural hearing aid, which are complemented with electroencephalography (EEG) recordings to infer the speaker of interest. METHODS: In this study, we propose a modular processing flow that first extracts the two speech envelopes from the microphone recordings, then selects the attended speech envelope based on the EEG, and finally uses this envelope to inform a multichannel speech separation and denoising algorithm. RESULTS: Strong suppression of interfering (unattended) speech and background noise is achieved, while the attended speech is preserved. Furthermore, EEG-based auditory attention detection (AAD) is shown to be robust to the use of noisy speech signals. CONCLUSIONS: Our results show that AAD-based speaker extraction from microphone array recordings is feasible and robust, even in noisy acoustic environments, and without access to the clean speech signals to perform EEG-based AAD. SIGNIFICANCE: Current research on AAD always assumes the availability of the clean speech signals, which limits the applicability in real settings. We have extended this research to detect the attended speaker even when only microphone recordings with noisy speech mixtures are available. This is an enabling ingredient for new brain-computer interfaces and effective filtering schemes in neuro-steered hearing prostheses. Here, we provide a first proof of concept for EEG-informed attended speaker extraction and denoising.

Adaptive attention-driven speech enhancement for EEG-informed hearing prostheses.

State-of-the-art hearing prostheses are equipped with acoustic noise reduction algorithms to improve speech intelligibility. Currently, one of the major challenges is to perform acoustic noise reduction in so-called cocktail party scenarios with multiple speakers, in particular because it is difficult-if not impossible-for the algorithm to determine which are the target speaker(s) that should be enhanced, and which speaker(s) should be treated as interfering sources. Recently, it has been shown that electroencephalography (EEG) can be used to perform auditory attention detection, i.e., to detect to which speaker a subject is attending based on recordings of neural activity. In this paper, we combine such an EEG-based auditory attention detection (AAD) paradigm with an acoustic noise reduction algorithm based on the multi-channel Wiener filter (MWF), leading to a neuro-steered MWF. In particular, we analyze how the AAD accuracy affects the noise suppression performance of an adaptive MWF in a sliding-window implementation, where the user switches his attention between two speakers.