Machine Learning Based Classification of Resting-State fMRI Features Exemplified by Metabolic State (Hunger/Satiety)

Centrality nearest-neighbor projected-distance regression (C-NPDR) feature selection for correlation-based predictors with application to resting-state fMRI study of major depressive disorder

BACKGROUND

Nearest-neighbor projected-distance regression (NPDR) is a metric-based machine learning feature selection algorithm that uses distances between samples and projected differences between variables to identify variables or features that may interact to affect the prediction of complex outcomes. Typical tabular bioinformatics data consist of separate variables of interest, such as genes or proteins. In contrast, resting-state functional MRI (rs-fMRI) data are composed of time-series for brain regions of interest (ROIs) for each subject, and these within-brain time-series are typically transformed into correlations between pairs of ROIs. These pairs of variables of interest can then be used as inputs for feature selection or other machine learning methods. Straightforward feature selection would return the most significant pairs of ROIs; however, it would also be beneficial to know the importance of individual ROIs.

RESULTS

We extend NPDR to compute the importance of individual ROIs from correlation-based features. We introduce correlation-difference and centrality-based versions of NPDR. Centrality-based NPDR can be coupled with any centrality method and can be coupled with importance scores other than NPDR, such as random forest importance scores. We develop a new simulation method using random network theory to generate artificial correlation data predictors with variations in correlations that affect class prediction.

CONCLUSIONS

We compared feature selection methods based on detection of functional simulated ROIs, and we applied the new centrality NPDR approach to a resting-state fMRI study of major depressive disorder (MDD) participants and healthy controls. We determined that the areas of the brain that have the strongest network effect on MDD include the middle temporal gyrus, the inferior temporal gyrus, and the dorsal entorhinal cortex. The resulting feature selection and simulation approaches can be applied to other domains that use correlation-based features.

Classification of female MDD patients with and without suicidal ideation using resting-state functional magnetic resonance imaging and machine learning

Spontaneous blood oxygen level-dependent signals can be indirectly recorded in different brain regions with functional magnetic resonance imaging. In this study resting-state functional magnetic resonance imaging was used to measure the differences in connectivity and activation seen in major depressive disorder (MDD) patients with and without suicidal ideation and the control group. For our investigation, a brain atlas containing 116 regions of interest was used. We also used four voxel-based connectivity models, including degree centrality, the fractional amplitude of low-frequency fluctuations (fALFF), regional homogeneity, and voxel-mirrored Homotopic Connectivity. Feature selection was conducted using a sequential backward floating selection approach along with a Random Forest Classifier and Elastic Net. While all four models yield significant results, fALFF demonstrated higher accuracy rates in classifying the three groups. Further analysis revealed three features that demonstrated statistically significant differences between these three, resulting in a 90.00% accuracy rate. Prominent features identified from our analysis, with suicide ideation as the key variable, included the Superior frontal gyrus (dorsolateral and orbital parts), the median cingulate, and the paracingulate gyri. These areas are associated with the Central Executive Control Network (ECN), the Default Mode Network, and the ECN, respectively. Comparing the results of MDD patients with suicidal ideation to those without suicidal ideations suggests dysfunctions in decision-making ability, in MDD females suffering from suicidal tendencies. This may be related to a lack of inhibition or emotion regulation capability, which contributes to suicidal ideations.

Regulation of craving for real-time fMRI neurofeedback based on individual classification

In previous real-time functional magnetic resonance imaging neurofeedback (rtfMRI-NF) studies on smoking craving, the focus has been on within-region activity or between-region connectivity, neglecting the potential predictive utility of broader network activity. Moreover, there is debate over the use and relative predictive power of individual-specific and group-level classifiers. This study aims to further advance rtfMRI-NF for substance use disorders by using whole-brain rtfMRI-NF to assess smoking craving-related brain patterns, evaluate the performance of group-level or individual-level classification (n = 31) and evaluate the performance of an optimized classifier across repeated NF runs. Using real-time individual-level classifiers derived from whole-brain support vector machines, we found that classification accuracy between crave and no-crave conditions and between repeated NF runs increased across repeated runs at both individual and group levels. In addition, individual-level accuracy was significantly greater than group-level accuracy, highlighting the potential increased utility of an individually trained whole-brain classifier for volitional control over brain patterns to regulate smoking craving. This study provides evidence supporting the feasibility of using whole-brain rtfMRI-NF to modulate smoking craving-related brain responses and the potential for learning individual strategies through optimization across repeated feedback runs. This article is part of the theme issue 'Neurofeedback: new territories and neurocognitive mechanisms of endogenous neuromodulation'.

Abnormal degree centrality as a potential imaging biomarker for ischemic stroke: A resting-state functional magnetic resonance imaging study

OBJECTIVE

To explore degree centrality (DC) abnormalities in ischemic stroke patients and determine whether these abnormalities have potential value in understanding the pathological mechanisms of ischemic stroke patients.

METHODS

Sixteen ischemic stroke patients and 22 healthy controls (HCs) underwent resting state functional magnetic resonance imaging (rs-fMRI) scanning, and the resulting data were subjected to DC analysis. Then we conducted a correlation analysis between DC values and neuropsychological test scores, including Montreal Cognitive Assessment (MoCA) and Mini-Mental State Examination (MMSE). Finally, extracted the abnormal DC values of brain regions and defined them as features for support vector machine (SVM) analysis.

RESULTS

Compared with HCs, ischemic stroke patients showed increased DC in the bilateral supplementary motor area, and median cingulate and paracingulate gyri and decreased DC in the left postcentral gyrus, right calcarine fissure and surrounding cortex, lingual gyrus, and orbital parts of the right superior frontal gyrus and bilateral cuneus. Correlation analyses revealed that DC values in the right lingual gyrus, calcarine fissure and surrounding cortex, and orbital parts of the right superior frontal gyrus were positively correlated with the MMSE scores. The SVM classification of the DC values achieved an area under the curve (AUC) of 0.93, an accuracy of 89.47%.

CONCLUSION

Our research results indicate that ischemic stroke patients exhibit abnormalities in the global connectivity mechanisms and patterns of the brain network. These abnormal changes may provide neuroimaging evidence for stroke-related motor, visual, and cognitive impairments, contribute to a deeper comprehension of the underlying pathophysiological mechanisms implicated in ischemic stroke.

A novel wearable ERP-based BCI approach to explicate hunger necessity

This study aimed to design a Brain-Computer Interface system to detect people's hunger status. EEG signals were recorded in various scenarios to create a database. We extracted the time-domain and frequency-domain features from these signals and applied them to the inputs of various Machine Learning algorithms. We compared the classification performances and reached the best-performing algorithm. The highest success score of 97.62% was achieved using the Multilayer Perceptron Neural Network algorithm in Event-Related Potential analysis.

Comparison of the diagnostic accuracy of resting-state fMRI driven machine learning algorithms in the detection of mild cognitive impairment

Mild cognitive impairment (MCI) is a potential therapeutic window in the prevention of dementia; however, automated detection of early cognitive deterioration is an unresolved issue. The aim of our study was to compare various classification approaches to differentiate MCI patients from healthy controls, based on rs-fMRI data, using machine learning (ML) algorithms. Own dataset (from two centers) and ADNI database were used during the analysis. Three fMRI parameters were applied in five feature selection algorithms: local correlation, intrinsic connectivity, and fractional amplitude of low frequency fluctuations. Support vector machine (SVM) and random forest (RF) methods were applied for classification. We achieved a relatively wide range of 78-87% accuracy for the various feature selection methods with SVM combining the three rs-fMRI parameters. In the ADNI datasets case we can also see even 90% accuracy scores. RF provided a more harmonized result among the feature selection algorithms in both datasets with 80-84% accuracy for our local and 74-82% for the ADNI database. Despite some lower performance metrics of some algorithms, most of the results were positive and could be seen in two unrelated datasets which increase the validity of our methods. Our results highlight the potential of ML-based fMRI applications for automated diagnostic techniques to recognize MCI patients.

SenseHunger: Machine Learning Approach to Hunger Detection Using Wearable Sensors

The perception of hunger and satiety is of great importance to maintaining a healthy body weight and avoiding chronic diseases such as obesity, underweight, or deficiency syndromes due to malnutrition. There are a number of disease patterns, characterized by a chronic loss of this perception. To our best knowledge, hunger and satiety cannot be classified using non-invasive measurements. Aiming to develop an objective classification system, this paper presents a multimodal sensory system using associated signal processing and pattern recognition methods for hunger and satiety detection based on non-invasive monitoring. We used an Empatica E4 smartwatch, a RespiBan wearable device, and JINS MEME smart glasses to capture physiological signals from five healthy normal weight subjects inactively sitting on a chair in a state of hunger and satiety. After pre-processing the signals, we compared different feature extraction approaches, either based on manual feature engineering or deep feature learning. Comparative experiments were carried out to determine the most appropriate sensor channel, device, and classifier to reliably discriminate between hunger and satiety states. Our experiments showed that the most discriminative features come from three specific sensor modalities: Electrodermal Activity (EDA), infrared Thermopile (Tmp), and Blood Volume Pulse (BVP).

A machine-learning approach for predicting impaired consciousness in absence epilepsy

Behavior during 3-4 Hz spike-wave discharges (SWDs) in absence epilepsy can vary from obvious behavioral arrest to no detectible deficits. Knowing if behavior is impaired is crucial for clinical care but may be difficult to determine without specialized behavioral testing, often inaccessible in practice. We aimed to develop a pure electroencephalography (EEG)-based machine-learning method to predict SWD-related behavioral impairment. Our classification goals were 100% predictive value, with no behaviorally impaired SWDs misclassified as spared; and maximal sensitivity. First, using labeled data with known behavior (130 SWDs in 34 patients), we extracted EEG time, frequency domain, and common spatial pattern features and applied support vector machines and linear discriminant analysis to classify SWDs as spared or impaired. We evaluated 32 classification models, optimized with 10-fold cross-validation. We then generalized these models to unlabeled data (220 SWDs in 41 patients), where behavior during individual SWDs was not known, but observers reported the presence of clinical seizures. For labeled data, the best classifier achieved 100% spared predictive value and 93% sensitivity. The best classifier on the unlabeled data achieved 100% spared predictive value, but with a lower sensitivity of 35%, corresponding to a conservative classification of 8 patients out of 23 as free of clinical seizures. Our findings demonstrate the feasibility of machine learning to predict impaired behavior during SWDs based on EEG features. With additional validation and optimization in a larger data sample, applications may include EEG-based prediction of driving safety, treatment adjustment, and insight into mechanisms of impaired consciousness in absence seizures.

Left DLPFC activity is associated with plasma kynurenine levels and can predict treatment response to escitalopram in major depressive disorder

AIM

To establish treatment response biomarkers that reflect the pathophysiology of depression, it is important to use an integrated set of features. This study aimed to determine the relationship between regional brain activity at rest and blood metabolites related to treatment response to escitalopram to identify the characteristics of depression that respond to treatment.

METHODS

Blood metabolite levels and resting-state brain activity were measured in patients with moderate to severe depression (n = 65) before and after 6-8 weeks of treatment with escitalopram, and these were compared between Responders and Nonresponders to treatment. We then examined the relationship between blood metabolites and brain activity related to treatment responsiveness in patients and healthy controls (n = 36).

RESULTS

Thirty-two patients (49.2%) showed a clinical response (>50% reduction in the Hamilton Rating Scale for Depression score) and were classified as Responders, and the remaining 33 patients were classified as Nonresponders. The pretreatment fractional amplitude of low-frequency fluctuation (fALFF) value of the left dorsolateral prefrontal cortex (DLPFC) and plasma kynurenine levels were lower in Responders, and the rate of increase of both after treatment was correlated with an improvement in symptoms. Moreover, the fALFF value of the left DLPFC was significantly correlated with plasma kynurenine levels in pretreatment patients with depression and healthy controls.

CONCLUSION

Decreased resting-state regional activity of the left DLPFC and decreased plasma kynurenine levels may predict treatment response to escitalopram, suggesting that it may be involved in the pathophysiology of major depressive disorder in response to escitalopram treatment.

Five weeks of intermittent transcutaneous vagus nerve stimulation shape neural networks: a machine learning approach

Invasive and transcutaneous vagus nerve stimulation [(t)-VNS] have been used to treat epilepsy, depression and migraine and has also shown effects on metabolism and body weight. To what extent this treatment shapes neural networks and how such network changes might be related to treatment effects is currently unclear. Using a pre-post mixed study design, we applied either a tVNS or sham stimulation (5 h/week) in 34 overweight male participants in the context of a study designed to assess effects of tVNS on body weight and metabolic and cognitive parameters resting state (rs) fMRI was measured about 12 h after the last stimulation period. Support vector machine (SVM) classification was applied to fractional amplitude low-frequency fluctuations (fALFF) on established rs-networks. All classification results were controlled for random effects and overfitting. Finally, we calculated multiple regressions between the classification results and reported food craving. We found a classification accuracy (CA) of 79 % in a subset of four brainstem regions suggesting that tVNS leads to lasting changes in brain networks. Five of eight salience network regions yielded 76,5 % CA. Our study shows tVNS’ post-stimulation effects on fALFF in the salience rs-network. More detailed investigations of this effect and their relationship with food intake seem reasonable for future studies.

A 3D multiscale view convolutional neural network with attention for mental disease diagnosis on MRI images

Computer Assisted Diagnosis (CAD) based on brain Magnetic Resonance Imaging (MRI) is a popular research field for the computer science and medical engineering. Traditional machine learning and deep learning methods were employed in the classification of brain MRI images in the previous studies. However, the current algorithms rarely take into consideration the influence of multi-scale brain connectivity disorders on some mental diseases. To improve this defect, a deep learning structure was proposed based on MRI images, which was designed to consider the brain's connections at different sizes and the attention of connections. In this work, a Multiscale View (MV) module was proposed, which was designed to detect multi-scale brain network disorders. On the basis of the MV module, the path attention module was also proposed to simulate the attention selection of the parallel paths in the MV module. Based on the two modules, we proposed a 3D Multiscale View Convolutional Neural Network with Attention (3D MVA-CNN) for classification of MRI images for mental disease. The proposed method outperformed the previous 3D CNN structures in the structural MRI data of ADHD-200 and the functional MRI data of schizophrenia. Finally, we also proposed a preliminary framework for clinical application using 3D CNN, and discussed its limitations on data accessing and reliability. This work promoted the assisted diagnosis of mental diseases based on deep learning and provided a novel 3D CNN method based on MRI data.

Single-participant structural similarity matrices lead to greater accuracy in classification of participants than function in autism in MRI

BACKGROUND

Autism has previously been characterized by both structural and functional differences in brain connectivity. However, while the literature on single-subject derivations of functional connectivity is extensively developed, similar methods of structural connectivity or similarity derivation from T1 MRI are less studied.

METHODS

We introduce a technique of deriving symmetric similarity matrices from regional histograms of grey matter volumes estimated from T1-weighted MRIs. We then validated the technique by inputting the similarity matrices into a convolutional neural network (CNN) to classify between participants with autism and age-, motion-, and intracranial-volume-matched controls from six different databases (29,288 total connectomes, mean age = 30.72, range 0.42-78.00, including 1555 subjects with autism). We compared this method to similar classifications of the same participants using fMRI connectivity matrices as well as univariate estimates of grey matter volumes. We further applied graph-theoretical metrics on output class activation maps to identify areas of the matrices that the CNN preferentially used to make the classification, focusing particularly on hubs.

LIMITATIONS

While this study used a large sample size, the majority of data was from a young age group; furthermore, to make a viable machine learning study, we treated autism, a highly heterogeneous condition, as a binary label. Thus, these results are not necessarily generalizable to all subtypes and age groups in autism.

RESULTS

Our models gave AUROCs of 0.7298 (69.71% accuracy) when classifying by only structural similarity, 0.6964 (67.72% accuracy) when classifying by only functional connectivity, and 0.7037 (66.43% accuracy) when classifying by univariate grey matter volumes. Combining structural similarity and functional connectivity gave an AUROC of 0.7354 (69.40% accuracy). Analysis of classification performance across age revealed the greatest accuracy in adolescents, in which most data were present. Graph analysis of class activation maps revealed no distinguishable network patterns for functional inputs, but did reveal localized differences between groups in bilateral Heschl's gyrus and upper vermis for structural similarity.

CONCLUSION

This study provides a simple means of feature extraction for inputting large numbers of structural MRIs into machine learning models. Our methods revealed a unique emphasis of the deep learning model on the structure of the bilateral Heschl's gyrus when characterizing autism.

Vector Phase Analysis Approach for Sleep Stage Classification: A Functional Near-Infrared Spectroscopy-Based Passive Brain-Computer Interface

A passive brain-computer interface (BCI) based upon functional near-infrared spectroscopy (fNIRS) brain signals is used for earlier detection of human drowsiness during driving tasks. This BCI modality acquired hemodynamic signals of 13 healthy subjects from the right dorsolateral prefrontal cortex (DPFC) of the brain. Drowsiness activity is recorded using a continuous-wave fNIRS system and eight channels over the right DPFC. During the experiment, sleep-deprived subjects drove a vehicle in a driving simulator while their cerebral oxygen regulation (CORE) state was continuously measured. Vector phase analysis (VPA) was used as a classifier to detect drowsiness state along with sleep stage-based threshold criteria. Extensive training and testing with various feature sets and classifiers are done to justify the adaptation of threshold criteria for any subject without requiring recalibration. Three statistical features (mean oxyhemoglobin, signal peak, and the sum of peaks) along with six VPA features (trajectory slopes of VPA indices) were used. The average accuracies for the five classifiers are 90.9% for discriminant analysis, 92.5% for support vector machines, 92.3% for nearest neighbors, 92.4% for both decision trees, and ensembles over all subjects' data. Trajectory slopes of CORE vector magnitude and angle: m(|R|) and m(∠R) are the best-performing features, along with ensemble classifier with the highest accuracy of 95.3% and minimum computation time of 40 ms. The statistical significance of the results is validated with a p-value of less than 0.05. The proposed passive BCI scheme demonstrates a promising technique for online drowsiness detection using VPA along with sleep stage classification.

Artificial intelligence for brain diseases: A systematic review

Artificial intelligence (AI) is a major branch of computer science that is fruitfully used for analyzing complex medical data and extracting meaningful relationships in datasets, for several clinical aims. Specifically, in the brain care domain, several innovative approaches have achieved remarkable results and open new perspectives in terms of diagnosis, planning, and outcome prediction. In this work, we present an overview of different artificial intelligent techniques used in the brain care domain, along with a review of important clinical applications. A systematic and careful literature search in major databases such as Pubmed, Scopus, and Web of Science was carried out using "artificial intelligence" and "brain" as main keywords. Further references were integrated by cross-referencing from key articles. 155 studies out of 2696 were identified, which actually made use of AI algorithms for different purposes (diagnosis, surgical treatment, intra-operative assistance, and postoperative assessment). Artificial neural networks have risen to prominent positions among the most widely used analytical tools. Classic machine learning approaches such as support vector machine and random forest are still widely used. Task-specific algorithms are designed for solving specific problems. Brain images are one of the most used data types. AI has the possibility to improve clinicians' decision-making ability in neuroscience applications. However, major issues still need to be addressed for a better practical use of AI in the brain. To this aim, it is important to both gather comprehensive data and build explainable AI algorithms.

Current Challenges in Translational and Clinical fMRI and Future Directions

Translational neuroscience is an important field that brings together clinical praxis with neuroscience methods. In this review article, the focus will be on functional neuroimaging (fMRI) and its applicability in clinical fMRI studies. In the light of the "replication crisis," three aspects will be critically discussed: First, the fMRI signal itself, second, current fMRI praxis, and, third, the next generation of analysis strategies. Current attempts such as resting-state fMRI, meta-analyses, and machine learning will be discussed with their advantages and potential pitfalls and disadvantages. One major concern is that the fMRI signal shows substantial within- and between-subject variability, which affects the reliability of both task-related, but in particularly resting-state fMRI studies. Furthermore, the lack of standardized acquisition and analysis methods hinders the further development of clinical relevant approaches. However, meta-analyses and machine-learning approaches may help to overcome current shortcomings in the methods by identifying new, and yet hidden relationships, and may help to build new models on disorder mechanisms. Furthermore, better control of parameters that may have an influence on the fMRI signal and that can easily be controlled for, like blood pressure, heart rate, diet, time of day, might improve reliability substantially.