The Diagnostic Accuracy of Artificial Intelligence in Radiological Markers of Normal-Pressure Hydrocephalus (NPH) on Non-Contrast CT Scans of the Brain.

Diagnosing normal-pressure hydrocephalus (NPH) via non-contrast computed tomography (CT) brain scans is presently a formidable task due to the lack of universally agreed-upon standards for radiographic parameter measurement. A variety of radiological parameters, such as Evans' index, narrow sulci at high parietal convexity, Sylvian fissures' dilation, focally enlarged sulci, and more, are currently measured by radiologists. This study aimed to enhance NPH diagnosis by comparing the accuracy, sensitivity, specificity, and predictive values of radiological parameters, as evaluated by radiologists and AI methods, utilizing cerebrospinal fluid volumetry. Results revealed a sensitivity of 77.14% for radiologists and 99.05% for AI, with specificities of 98.21% and 57.14%, respectively, in diagnosing NPH. Radiologists demonstrated NPV, PPV, and an accuracy of 82.09%, 97.59%, and 88.02%, while AI reported 98.46%, 68.42%, and 77.42%, respectively. ROC curves exhibited an area under the curve of 0.954 for radiologists and 0.784 for AI, signifying the diagnostic index for NPH. In conclusion, although radiologists exhibited superior sensitivity, specificity, and accuracy in diagnosing NPH, AI served as an effective initial screening mechanism for potential NPH cases, potentially easing the radiologists' burden. Given the ongoing AI advancements, it is plausible that AI could eventually match or exceed radiologists' diagnostic prowess in identifying hydrocephalus.

Quantifying the impact of data characteristics on the transferability of sleep stage scoring models.

Deep learning models for scoring sleep stages based on single-channel EEG have been proposed as a promising method for remote sleep monitoring. However, applying these models to new datasets, particularly from wearable devices, raises two questions. First, when annotations on a target dataset are unavailable, which different data characteristics affect the sleep stage scoring performance the most and by how much? Second, when annotations are available, which dataset should be used as the source of transfer learning to optimize performance? In this paper, we propose a novel method for computationally quantifying the impact of different data characteristics on the transferability of deep learning models. Quantification is accomplished by training and evaluating two models with significant architectural differences, TinySleepNet and U-Time, under various transfer configurations in which the source and target datasets have different recording channels, recording environments, and subject conditions. For the first question, the environment had the highest impact on sleep stage scoring performance, with performance degrading by over 14% when sleep annotations were unavailable. For the second question, the most useful transfer sources for TinySleepNet and the U-Time models were MASS-SS1 and ISRUC-SG1, containing a high percentage of N1 (the rarest sleep stage) relative to the others. The frontal and central EEGs were preferred for TinySleepNet. The proposed approach enables full utilization of existing sleep datasets for training and planning model transfer to maximize the sleep stage scoring performance on a target problem when sleep annotations are limited or unavailable, supporting the realization of remote sleep monitoring.

TinySleepNet: An Efficient Deep Learning Model for Sleep Stage Scoring based on Raw Single-Channel EEG.

Deep learning has become popular for automatic sleep stage scoring due to its capability to extract useful features from raw signals. Most of the existing models, however, have been overengineered to consist of many layers or have introduced additional steps in the processing pipeline, such as converting signals to spectrogram-based images. They require to be trained on a large dataset to prevent the overfitting problem (but most of the sleep datasets contain a limited amount of class-imbalanced data) and are difficult to be applied (as there are many hyperparameters to be configured in the pipeline). In this paper, we propose an efficient deep learning model, named TinySleepNet, and a novel technique to effectively train the model end-to-end for automatic sleep stage scoring based on raw single-channel EEG. Our model consists of a less number of model parameters to be trained compared to the existing ones, requiring a less amount of training data and computational resources. Our training technique incorporates data augmentation that can make our model be more robust the shift along the time axis, and can prevent the model from remembering the sequence of sleep stages. We evaluated our model on seven public sleep datasets that have different characteristics in terms of scoring criteria and recording channels and environments. The results show that, with the same model architecture and the training parameters, our method achieves a similar (or better) performance compared to the state-of-the-art methods on all datasets. This demonstrates that our method can generalize well to the largest number of different datasets.

Remote Monitoring in the Home Validates Clinical Gait Measures for Multiple Sclerosis.

Background: The timed 25-foot walk (T25FW) is widely used as a clinic performance measure, but has yet to be directly validated against gait speed in the home environment. Objectives: To develop an accurate method for remote assessment of walking speed and to test how predictive the clinic T25FW is for real-life walking. Methods: An AX3-Axivity tri-axial accelerometer was positioned on 32 MS patients (Expanded Disability Status Scale [EDSS] 0-6) in the clinic, who subsequently wore it at home for up to 7 days. Gait speed was calculated from these data using both a model developed with healthy volunteers and individually personalized models generated from a machine learning algorithm. Results: The healthy volunteer model predicted gait speed poorly for more disabled people with MS. However, the accuracy of individually personalized models was high regardless of disability (R-value = 0.98, p-value = 1.85 × 10-22). With the latter, we confirmed that the clinic T25FW is strongly predictive of the maximum sustained gait speed in the home environment (R-value = 0.89, p-value = 4.34 × 10-8). Conclusion: Remote gait monitoring with individually personalized models is accurate for patients with MS. Using these models, we have directly validated the clinical meaningfulness (i.e., predictiveness) of the clinic T25FW for the first time.

Using Support Vector Machine on EEG for Advertisement Impact Assessment.

The advertising industry depends on an effective assessment of the impact of advertising as a key performance metric for their products. However, current assessment methods have relied on either indirect inference from observing changes in consumer behavior after the launch of an advertising campaign, which has long cycle times and requires an ad campaign to have already have been launched (often meaning costs having been sunk). Or through surveys or focus groups, which have a potential for experimental biases, peer pressure, and other psychological and sociological phenomena that can reduce the effectiveness of the study. In this paper, we investigate a new approach to assess the impact of advertisement by utilizing low-cost EEG headbands to record and assess the measurable impact of advertising on the brain. Our evaluation shows the desired performance of our method based on user experiment with 30 recruited subjects after watching 220 different advertisements. We believe the proposed SVM method can be further developed to a general and scalable methodology that can enable advertising agencies to assess impact rapidly, quantitatively, and without bias.

Mixed Neural Network Approach for Temporal Sleep Stage Classification.

This paper proposes a practical approach to addressing limitations posed by using of single-channel electroencephalography (EEG) for sleep stage classification. EEG-based characterizations of sleep stage progression contribute the diagnosis and monitoring of the many pathologies of sleep. Several prior reports explored ways of automating the analysis of sleep EEG and of reducing the complexity of the data needed for reliable discrimination of sleep stages at lower cost in the home. However, these reports have involved recordings from electrodes placed on the cranial vertex or occiput, which are both uncomfortable and difficult to position. Previous studies of sleep stage scoring that used only frontal electrodes with a hierarchical decision tree motivated this paper, in which we have taken advantage of rectifier neural network for detecting hierarchical features and long short-term memory network for sequential data learning to optimize classification performance with single-channel recordings. After exploring alternative electrode placements, we found a comfortable configuration of a single-channel EEG on the forehead and have shown that it can be integrated with additional electrodes for simultaneous recording of the electro-oculogram. Evaluation of data from 62 people (with 494 hours sleep) demonstrated better performance of our analytical algorithm than is available from existing approaches with vertex or occipital electrode placements. Use of this recording configuration with neural network deconvolution promises to make clinically indicated home sleep studies practical.

DeepSleepNet: A Model for Automatic Sleep Stage Scoring Based on Raw Single-Channel EEG.

This paper proposes a deep learning model, named DeepSleepNet, for automatic sleep stage scoring based on raw single-channel EEG. Most of the existing methods rely on hand-engineered features, which require prior knowledge of sleep analysis. Only a few of them encode the temporal information, such as transition rules, which is important for identifying the next sleep stages, into the extracted features. In the proposed model, we utilize convolutional neural networks to extract time-invariant features, and bidirectional-long short-term memory to learn transition rules among sleep stages automatically from EEG epochs. We implement a two-step training algorithm to train our model efficiently. We evaluated our model using different single-channel EEGs (F4-EOG (left), Fpz-Cz, and Pz-Oz) from two public sleep data sets, that have different properties (e.g., sampling rate) and scoring standards (AASM and R&K). The results showed that our model achieved similar overall accuracy and macro F1-score (MASS: 86.2%-81.7, Sleep-EDF: 82.0%-76.9) compared with the state-of-the-art methods (MASS: 85.9%-80.5, Sleep-EDF: 78.9%-73.7) on both data sets. This demonstrated that, without changing the model architecture and the training algorithm, our model could automatically learn features for sleep stage scoring from different raw single-channel EEGs from different data sets without utilizing any hand-engineered features.

Feature extraction with stacked autoencoders for epileptic seizure detection.

Scalp electroencephalogram (EEG), a recording of the brain's electrical activity, has been used to diagnose and detect epileptic seizures for a long time. However, most researchers have implemented seizure detectors by manually hand-engineering features from observed EEG data, and used them in seizure detection, which might not scale well to new patterns of seizures. In this paper, we investigate the possibility of utilising unsupervised feature learning, the recent development of deep learning, to automatically learn features from raw, unlabelled EEG data that are representative enough to be used in seizure detection. We develop patient-specific seizure detectors by using stacked autoencoders and logistic classifiers. A two-step training consisting of the greedy layer-wise and the global fine-tuning was used to train our detectors. The evaluation was performed by using labelled dataset from the CHB-MIT database, and the results showed that all of the test seizures were detected with a mean latency of 3.36 seconds, and a low false detection rate.