Single subject prediction of brain disorders in neuroimaging: Promises and pitfalls

Altered network stability in progressive supranuclear palsy

•

We investigated network dynamics in the tauopathy progressive supranuclear palsy

•

Abnormal temporal properties of large-scale networks are related to phenotype

•

Progressive supranuclear palsy paradoxically increases frontoparietal state time

•

Reductions in neural signal complexity relate to altered network dynamics

•

Dynamic network and topological changes occur distally to primary sites of atrophy

The clinical syndromes of Progressive Supranuclear Palsy (PSP) may be mediated by abnormal temporal dynamics of brain networks, due to the impact of atrophy, synapse loss and neurotransmitter deficits. We tested the hypothesis that alterations in signal complexity in neural networks influence short-latency state transitions. Ninety-four participants with PSP and 64 healthy controls were recruited from two independent cohorts. All participants underwent clinical and neuropsychological testing and resting-state functional MRI. Network dynamics were assessed using hidden Markov models and neural signal complexity measured in terms of multiscale entropy. In both cohorts, PSP increased the proportion of time in networks associated with higher cognitive functions. This effect correlated with clinical severity as measured by the PSP-rating-scale, and with reduced neural signal complexity. Regional atrophy influenced abnormal brain-state occupancy, but abnormal network topology and dynamics were not restricted to areas of atrophy. Our findings show that the pathology of PSP causes clinically relevant changes in neural temporal dynamics, leading to a greater proportion of time in inefficient brain-states.

Mixed-effects multilevel analysis followed by canonical correlation analysis is an effective fMRI tool for the investigation of idiosyncrasies

We report that regions-of-interest (ROIs) associated with idiosyncratic individual behavior can be identified from functional magnetic resonance imaging (fMRI) data using statistical approaches that explicitly model individual variability in neuronal activations, such as mixed-effects multilevel analysis (MEMA). We also show that the relationship between neuronal activation in fMRI and behavioral data can be modeled using canonical correlation analysis (CCA). A real-world dataset for the neuronal response to nicotine use was acquired using a custom-made MRI-compatible apparatus for the smoking of electronic cigarettes (e-cigarettes). Nineteen participants smoked e-cigarettes in an MRI scanner using the apparatus with two experimental conditions: e-cigarettes with nicotine (ECIG) and sham e-cigarettes without nicotine (SCIG) and subjective ratings were collected. The right insula was identified in the ECIG condition from the χ² -test of the MEMA but not from the t-test, and the corresponding activations were significantly associated with the similarity scores (r = -.52, p = .041, confidence interval [CI] = [-0.78, -0.17]) and the urge-to-smoke scores (r = .73, p <.001, CI = [0.52, 0.88]). From the contrast between the two conditions (i.e., ECIG > SCIG), the right orbitofrontal cortex was identified from the χ² -tests, and the corresponding neuronal activations showed a statistically meaningful association with similarity (r = -.58, p = .01, CI = [-0.84, -0.17]) and the urge to smoke (r = .34, p = .15, CI = [0.09, 0.56]). The validity of our analysis pipeline (i.e., MEMA followed by CCA) was further evaluated using the fMRI and behavioral data acquired from the working memory and gambling tasks available from the Human Connectome Project.

Mapping individual differences across brain network structure to function and behavior with connectome embedding

The connectome, a comprehensive map of the brain’s anatomical connections, is often summarized as a matrix comprising all dyadic connections among pairs of brain regions. This representation cannot capture higher-order relations within the brain graph. Connectome embedding (CE) addresses this limitation by creating compact vectorized representations of brain nodes capturing their context in the global network topology. Here, nodes “context” is defined as random walks on the brain graph and as such, represents a generative model of diffusive communication around nodes. Applied to group-averaged structural connectivity, CE was previously shown to capture relations between inter-hemispheric homologous brain regions and uncover putative missing edges from the network reconstruction. Here we extend this framework to explore individual differences with a novel embedding alignment approach. We test this approach in two lifespan datasets (NKI: n = 542; Cam-CAN: n = 601) that include diffusion-weighted imaging, resting-state fMRI, demographics and behavioral measures. We demonstrate that modeling functional connectivity with CE substantially improves structural to functional connectivity mapping both at the group and subject level. Furthermore, age-related differences in this structure-function mapping, are preserved and enhanced. Importantly, CE captures individual differences by out-of-sample prediction of age and intelligence. The resulting predictive accuracy was higher compared to using structural connectivity and functional connectivity. We attribute these findings to the capacity of the CE to incorporate aspects of both anatomy (the structural graph) and function (diffusive communication). Our novel approach allows mapping individual differences in the connectome through structure to function and behavior.

Data-driven approaches to neuroimaging biomarkers for neurological and psychiatric disorders: emerging approaches and examples

PURPOSE OF REVIEW

The 'holy grail' of clinical applications of neuroimaging to neurological and psychiatric disorders via personalized biomarkers has remained mostly elusive, despite considerable effort. However, there are many reasons to continue to be hopeful, as the field has made remarkable advances over the past few years, fueled by a variety of converging technical and data developments.

RECENT FINDINGS

We discuss a number of advances that are accelerating the push for neuroimaging biomarkers including the advent of the 'neuroscience big data' era, biomarker data competitions, the development of more sophisticated algorithms including 'guided' data-driven approaches that facilitate automation of network-based analyses, dynamic connectivity, and deep learning. Another key advance includes multimodal data fusion approaches which can provide convergent and complementary evidence pointing to possible mechanisms as well as increase predictive accuracy.

SUMMARY

The search for clinically relevant neuroimaging biomarkers for neurological and psychiatric disorders is rapidly accelerating. Here, we highlight some of these aspects, provide recent examples from studies in our group, and link to other ongoing work in the field. It is critical that access and use of these advanced approaches becomes mainstream, this will help propel the community forward and facilitate the production of robust and replicable neuroimaging biomarkers.

Artificial Intelligence for Hospital Health Care: Application Cases and Answers to Challenges in European Hospitals

The development and implementation of artificial intelligence (AI) applications in health care contexts is a concurrent research and management question. Especially for hospitals, the expectations regarding improved efficiency and effectiveness by the introduction of novel AI applications are huge. However, experiences with real-life AI use cases are still scarce. As a first step towards structuring and comparing such experiences, this paper is presenting a comparative approach from nine European hospitals and eleven different use cases with possible application areas and benefits of hospital AI technologies. This is structured as a current review and opinion article from a diverse range of researchers and health care professionals. This contributes to important improvement options also for pandemic crises challenges, e.g., the current COVID-19 situation. The expected advantages as well as challenges regarding data protection, privacy, or human acceptance are reported. Altogether, the diversity of application cases is a core characteristic of AI applications in hospitals, and this requires a specific approach for successful implementation in the health care sector. This can include specialized solutions for hospitals regarding human-computer interaction, data management, and communication in AI implementation projects.

Predicting subclinical psychotic-like experiences on a continuum using machine learning

Previous studies applying machine learning methods to psychosis have primarily been concerned with the binary classification of chronic schizophrenia patients and healthy controls. The aim of this study was to use electroencephalographic (EEG) data and pattern recognition to predict subclinical psychotic-like experiences on a continuum between these two extremes in otherwise healthy people. We applied two different approaches to an auditory oddball regularity learning task obtained from N = 73 participants: A feature extraction and selection routine incorporating behavioural measures, event-related potential components and effective connectivity parameters; Regularisation of spatiotemporal maps of event-related potentials. Using the latter approach, optimal performance was achieved using the response to frequent, predictable sounds. Features within the P50 and P200 time windows had the greatest contribution toward lower Prodromal Questionnaire (PQ) scores and the N100 time window contributed most to higher PQ scores. As a proof-of-concept, these findings demonstrate that EEG data alone are predictive of individual psychotic-like experiences in healthy people. Our findings are in keeping with the mounting evidence for altered sensory responses in schizophrenia, as well as the notion that psychosis may exist on a continuum expanding into the non-clinical population.

Functional Disruptions of the Brain in Low Back Pain: A Potential Imaging Biomarker of Functional Disability

Chronic low back pain (LBP) is one of the leading causes of disability worldwide. While LBP research has largely focused on the spine, many studies have demonstrated a restructuring of human brain architecture accompanying LBP and other chronic pain states. Brain imaging presents a promising source for discovering noninvasive biomarkers that can improve diagnostic and prognostication outcomes for chronic LBP. This study evaluated graph theory measures derived from brain resting-state functional connectivity (rsFC) as prospective noninvasive biomarkers of LBP. We also proposed and tested a hybrid feature selection method (Enet-subset) that combines Elastic Net and an optimal subset selection method. We collected resting-state functional MRI scans from 24 LBP patients and 27 age-matched healthy controls (HC). We then derived graph-theoretical features and trained a support vector machine (SVM) to classify patient group. The degree centrality (DC), clustering coefficient (CC), and betweenness centrality (BC) were found to be significant predictors of patient group. We achieved an average classification accuracy of 83.1% (p < 0.004) and AUC of 0.937 (p < 0.002), respectively. Similarly, we achieved a sensitivity and specificity of 87.0 and 79.7%. The classification results from this study suggest that graph matrices derived from rsFC can be used as biomarkers of LBP. In addition, our findings suggest that the proposed feature selection method, Enet-subset, might act as a better technique to remove redundant variables and improve the performance of the machine learning classifier.

Systematic misestimation of machine learning performance in neuroimaging studies of depression

We currently observe a disconcerting phenomenon in machine learning studies in psychiatry: While we would expect larger samples to yield better results due to the availability of more data, larger machine learning studies consistently show much weaker performance than the numerous small-scale studies. Here, we systematically investigated this effect focusing on one of the most heavily studied questions in the field, namely the classification of patients suffering from Major Depressive Disorder (MDD) and healthy controls based on neuroimaging data. Drawing upon structural MRI data from a balanced sample of N = 1868 MDD patients and healthy controls from our recent international Predictive Analytics Competition (PAC), we first trained and tested a classification model on the full dataset which yielded an accuracy of 61%. Next, we mimicked the process by which researchers would draw samples of various sizes (N = 4 to N = 150) from the population and showed a strong risk of misestimation. Specifically, for small sample sizes (N = 20), we observe accuracies of up to 95%. For medium sample sizes (N = 100) accuracies up to 75% were found. Importantly, further investigation showed that sufficiently large test sets effectively protect against performance misestimation whereas larger datasets per se do not. While these results question the validity of a substantial part of the current literature, we outline the relatively low-cost remedy of larger test sets, which is readily available in most cases.

Identifying Autism Spectrum Disorder From Resting-State fMRI Using Deep Belief Network

With the increasing prevalence of autism spectrum disorder (ASD), it is important to identify ASD patients for effective treatment and intervention, especially in early childhood. Neuroimaging techniques have been used to characterize the complex biomarkers based on the functional connectivity anomalies in the ASD. However, the diagnosis of ASD still adopts the symptom-based criteria by clinical observation. The existing computational models tend to achieve unreliable diagnostic classification on the large-scale aggregated data sets. In this work, we propose a novel graph-based classification model using the deep belief network (DBN) and the Autism Brain Imaging Data Exchange (ABIDE) database, which is a worldwide multisite functional and structural brain imaging data aggregation. The remarkable connectivity features are selected through a graph extension of K -nearest neighbors and then refined by a restricted path-based depth-first search algorithm. Thanks to the feature reduction, lower computational complexity could contribute to the shortening of the training time. The automatic hyperparameter-tuning technique is introduced to optimize the hyperparameters of the DBN by exploring the potential parameter space. The simulation experiments demonstrate the superior performance of our model, which is 6.4% higher than the best result reported on the ABIDE database. We also propose to use the data augmentation and the oversampling technique to identify further the possible subtypes within the ASD. The interpretability of our model enables the identification of the most remarkable autistic neural correlation patterns from the data-driven outcomes.

Personalized prediction of transcranial magnetic stimulation clinical response in patients with treatment-refractory depression using neuroimaging biomarkers and machine learning

BACKGROUND

Functional connectivity between the left dorsolateral prefrontal cortex (DLPFC) and subgenual cingulate (sgACC) may serve as a biomarker for transcranial magnetic stimulation (rTMS) treatment response. The first aim was to establish whether this finding is veridical or artifactually induced by the pre-processing method. Furthermore, alternative biomarkers were identified and the clinical utility for personalized medicine was examined.

METHODS

Resting-state fMRI data were collected in medication-refractory depressed patients (n = 70, 16 males) before undergoing neuronavigated left DLPFC rTMS. Seed-based analyses were performed with and without global signal regression pre-processing to identify biomarkers of short-term and long-term treatment response. Receiver Operating Characteristic curve and supervised machine learning analyses were applied to assess the clinical utility of these biomarkers for the classification of categorical rTMS response.

RESULTS

Regardless of the pre-processing method, DLPFC-sgACC connectivity was not associated with treatment outcome. Instead, poorer connectivity between the sgACC and three clusters (peak locations: frontal pole, superior parietal lobule, occipital cortex) and DLPFC-central opercular cortex were observed in long-term nonresponders. The identified connections could serve as acceptable to excellent markers. Combining the features using supervised machine learning reached accuracy rates of 95.35% (CI=82.94-100.00) and 88.89% (CI=63.96-100.00) in the cross-validation and test dataset, respectively.

LIMITATIONS

The sample size was moderate, and features for machine learning were based on group differences.

CONCLUSIONS

Long-term nonresponders showed greater disrupted connectivity in regions involving the central executive network. Our findings may aid the development of personalized medicine for medication-refractory depression.

Distributed functional connectivity predicts neuropsychological test performance among older adults

Neuropsychological test is an essential tool in assessing cognitive and functional changes associated with late-life neurocognitive disorders. Despite the utility of the neuropsychological test, the brain-wide neural basis of the test performance remains unclear. Using the predictive modeling approach, we aimed to identify the optimal combination of functional connectivities that predicts neuropsychological test scores of novel individuals. Resting-state functional connectivity and neuropsychological tests included in the OASIS-3 dataset (n = 428) were used to train the predictive models, and the identified models were iteratively applied to the holdout internal test set (n = 216) and external test set (KSHAP, n = 151). We found that the connectivity-based predicted score tracked the actual behavioral test scores (r = 0.08-0.44). The predictive models utilizing most of the connectivity features showed better accuracy than those composed of focal connectivity features, suggesting that its neural basis is largely distributed across multiple brain systems. The discriminant and clinical validity of the predictive models were further assessed. Our results suggest that late-life neuropsychological test performance can be formally characterized with distributed connectome-based predictive models, and further translational evidence is needed when developing theoretically valid and clinically incremental predictive models.

Brain imaging-based machine learning in autism spectrum disorder: methods and applications

Autism spectrum disorder (ASD) is a neurodevelopmental condition with early childhood onset and high heterogeneity. As the pathogenesis is still elusive, ASD diagnosis is comprised of a constellation of behavioral symptoms. Non-invasive brain imaging techniques, such as magnetic resonance imaging (MRI), provide a valuable objective measurement of the brain. Many efforts have been devoted to developing imaging-based diagnostic tools for ASD based on machine learning (ML) technologies. In this survey, we review recent advances that utilize machine learning approaches to classify individuals with and without ASD. First, we provide a brief overview of neuroimaging-based ASD classification studies, including the analysis of publications and general classification pipeline. Next, representative studies are highlighted and discussed in detail regarding different imaging modalities, methods and sample sizes. Finally, we highlight several common challenges and provide recommendations on future directions. In summary, identifying discriminative biomarkers for ASD diagnosis is challenging, and further establishing more comprehensive datasets and dissecting the individual and group heterogeneity will be critical to achieve better ADS diagnosis performance. Machine learning methods will continue to be developed and are poised to help advance the field in this regard.

In vivo biomarkers of structural and functional brain development and aging in humans

Brain aging is a major determinant of aging. Along with the aging population, prevalence of neurodegenerative diseases is increasing, therewith placing economic and social burden on individuals and society. Individual rates of brain aging are shaped by genetics, epigenetics, and prenatal environmental. Biomarkers of biological brain aging are needed to predict individual trajectories of aging and the risk for age-associated neurological impairments for developing early preventive and interventional measures. We review current advances of in vivo biomarkers predicting individual brain age. Telomere length and epigenetic clock, two important biomarkers that are closely related to the mechanistic aging process, have only poor deterministic and predictive accuracy regarding individual brain aging due to their high intra- and interindividual variability. Phenotype-related biomarkers of global cognitive function and brain structure provide a much closer correlation to age at the individual level. During fetal and perinatal life, autonomic activity is a unique functional marker of brain development. The cognitive and structural biomarkers also boast high diagnostic specificity for determining individual risks for neurodegenerative diseases.

Identifying Internet Addiction and Evaluating the Efficacy of Treatment Based on Functional Connectivity Density: A Machine Learning Study

Although mounting neuroimaging studies have greatly improved our understanding of the neurobiological mechanism underlying internet addiction (IA), the results based on traditional group-level comparisons are insufficient in guiding individual clinical practice directly. Specific neuroimaging biomarkers are urgently needed for IA diagnosis and the evaluation of therapy efficacy. Therefore, this study aimed to develop support vector machine (SVM) models to identify IA and assess the efficacy of cognitive behavior therapy (CBT) based on unbiased functional connectivity density (FCD). Resting-state fMRI data were acquired from 27 individuals with IA before and after 8-week CBT sessions and 30 demographically matched healthy controls (HCs). The discriminative FCDs were computed as the features of the support vector classification (SVC) model to identify individuals with IA from HCs, and the changes in these discriminative FCDs after treatment were further used as features of the support vector regression (SVR) model to evaluate the efficacy of CBT. Based on the informative FCDs, our SVC model successfully differentiated individuals with IA from HCs with an accuracy of 82.5% and an area under the curve (AUC) of 0.91. Our SVR model successfully evaluated the efficacy of CBT using the FCD change ratio with a correlation efficient of 0.59. The brain regions contributing to IA classification and CBT efficacy assessment were the left inferior frontal cortex (IFC), middle frontal cortex (MFC) and angular gyrus (AG), the right premotor cortex (PMC) and middle cingulate cortex (MCC), and the bilateral cerebellum, orbitofrontal cortex (OFC) and superior frontal cortex (SFC). These findings confirmed the FCDs of hyperactive impulsive habit system, hypoactive reflecting system and sensitive interoceptive reward awareness system as potential neuroimaging biomarkers for IA, which might provide objective indexes for the diagnosis and efficacy evaluation of IA.

Narrowband Resting-State fNIRS Functional Connectivity in Autism Spectrum Disorder

Hemispheric asymmetry in the power spectrum of low-frequency spontaneous hemodynamic fluctuations has been previously observed in autism spectrum disorder (ASD). This observation may imply a specific narrow-frequency band in which individuals with ASD could show more significant alteration in resting-state functional connectivity (RSFC). To test this assumption, we evaluated narrowband RSFC at several frequencies for functional near-infrared spectroscopy signals recorded from the bilateral temporal lobes on 25 children with ASD and 22 typically developing (TD) children. In several narrow-frequency bands, we observed altered interhemispheric RSFC in ASD. However, in the band of 0.01–0.02 Hz, more mirrored channel pairs (or cortical sites) showed significantly weaker RSFC in the ASD group. Receiver operating characteristic analysis further demonstrated that RSFC in the narrowband of 0.01–0.02 Hz might have better differentiation ability between the ASD and TD groups. This may indicate that the narrowband RSFC could serve as a characteristic for the prediction of ASD.

Cross-cohort generalizability of deep and conventional machine learning for MRI-based diagnosis and prediction of Alzheimer's disease

This work validates the generalizability of MRI-based classification of Alzheimer’s disease (AD) patients and controls (CN) to an external data set and to the task of prediction of conversion to AD in individuals with mild cognitive impairment (MCI).

We used a conventional support vector machine (SVM) and a deep convolutional neural network (CNN) approach based on structural MRI scans that underwent either minimal pre-processing or more extensive pre-processing into modulated gray matter (GM) maps. Classifiers were optimized and evaluated using cross-validation in the Alzheimer’s Disease Neuroimaging Initiative (ADNI; 334 AD, 520 CN). Trained classifiers were subsequently applied to predict conversion to AD in ADNI MCI patients (231 converters, 628 non-converters) and in the independent Health-RI Parelsnoer Neurodegenerative Diseases Biobank data set. From this multi-center study representing a tertiary memory clinic population, we included 199 AD patients, 139 participants with subjective cognitive decline, 48 MCI patients converting to dementia, and 91 MCI patients who did not convert to dementia.

AD-CN classification based on modulated GM maps resulted in a similar area-under-the-curve (AUC) for SVM (0.940; 95%CI: 0.924–0.955) and CNN (0.933; 95%CI: 0.918–0.948). Application to conversion prediction in MCI yielded significantly higher performance for SVM (AUC = 0.756; 95%CI: 0.720-0.788) than for CNN (AUC = 0.742; 95%CI: 0.709-0.776) (p<0.01 for McNemar’s test). In external validation, performance was slightly decreased. For AD-CN, it again gave similar AUCs for SVM (0.896; 95%CI: 0.855–0.932) and CNN (0.876; 95%CI: 0.836–0.913). For prediction in MCI, performances decreased for both SVM (AUC = 0.665; 95%CI: 0.576-0.760) and CNN (AUC = 0.702; 95%CI: 0.624-0.786). Both with SVM and CNN, classification based on modulated GM maps significantly outperformed classification based on minimally processed images (p=0.01).

Deep and conventional classifiers performed equally well for AD classification and their performance decreased only slightly when applied to the external cohort. We expect that this work on external validation contributes towards translation of machine learning to clinical practice.

Predictors of training-related improvement in visuomotor performance in patients with multiple sclerosis: A behavioural and MRI study

BACKGROUND

The development of tailored recovery-oriented strategies in multiple sclerosis requires early identification of an individual's potential for functional recovery.

OBJECTIVE

To identify predictors of visuomotor performance improvements, a proxy of functional recovery, using a predictive statistical model that combines demographic, clinical and magnetic resonance imaging (MRI) data.

METHODS

Right-handed multiple sclerosis patients underwent baseline disability assessment and MRI of the brain structure, function and vascular health. They subsequently undertook 4 weeks of right upper limb visuomotor practice. Changes in performance with practice were our outcome measure. We identified predictors of improvement in a training set of patients using lasso regression; we calculated the best performing model in a validation set and applied this model to a test set.

RESULTS

Patients improved their visuomotor performance with practice. Younger age, better visuomotor abilities, less severe disease burden and concurrent use of preventive treatments predicted improvements. Neuroimaging localised outcome-relevant sensory motor regions, the microstructure and activity of which correlated with performance improvements.

CONCLUSION

Initial characteristics, including age, disease duration, visuo-spatial abilities, hand dexterity, self-evaluated disease impact and the presence of disease-modifying treatments, can predict functional recovery in individual patients, potentially improving their clinical management and stratification in clinical trials. MRI is a correlate of outcome, potentially supporting individual prognosis.

Classification of first-episode psychosis using cortical thickness: A large multicenter MRI study

Machine learning classifications of first-episode psychosis (FEP) using neuroimaging have predominantly analyzed brain volumes. Some studies examined cortical thickness, but most of them have used parcellation approaches with data from single sites, which limits claims of generalizability. To address these limitations, we conducted a large-scale, multi-site analysis of cortical thickness comparing parcellations and vertex-wise approaches. By leveraging the multi-site nature of the study, we further investigated how different demographical and site-dependent variables affected predictions. Finally, we assessed relationships between predictions and clinical variables. 428 subjects (147 females, mean age 27.14) with FEP and 448 (230 females, mean age 27.06) healthy controls were enrolled in 8 centers by the ClassiFEP group. All subjects underwent a structural MRI and were clinically assessed. Cortical thickness parcellation (68 areas) and full cortical maps (20,484 vertices) were extracted. Linear Support Vector Machine was used for classification within a repeated nested cross-validation framework. Vertex-wise thickness maps outperformed parcellation-based methods with a balanced accuracy of 66.2% and an Area Under the Curve of 72%. By stratifying our sample for MRI scanner, we increased generalizability across sites. Temporal brain areas resulted as the most influential in the classification. The predictive decision scores significantly correlated with age at onset, duration of treatment, and positive symptoms. In conclusion, although far from the threshold of clinical relevance, temporal cortical thickness proved to classify between FEP subjects and healthy individuals. The assessment of site-dependent variables permitted an increase in the across-site generalizability, thus attempting to address an important machine learning limitation.

FMRI hemodynamic response function (HRF) as a novel marker of brain function: applications for understanding obsessive-compulsive disorder pathology and treatment response

The hemodynamic response function (HRF) represents the transfer function linking neural activity with the functional MRI (fMRI) signal, modeling neurovascular coupling. Since HRF is influenced by non-neural factors, to date it has largely been considered as a confound or has been ignored in many analyses. However, underlying biophysics suggests that the HRF may contain meaningful correlates of neural activity, which might be unavailable through conventional fMRI metrics. Here, we estimated the HRF by performing deconvolution on resting-state fMRI data from a longitudinal sample of 25 healthy controls scanned twice and 44 adults with obsessive-compulsive disorder (OCD) before and after 4-weeks of intensive cognitive-behavioral therapy (CBT). HRF response height, time-to-peak and full-width at half-maximum (FWHM) in OCD were abnormal before treatment and normalized after treatment in regions including the caudate. Pre-treatment HRF predicted treatment outcome (OCD symptom reduction) with 86.4% accuracy, using machine learning. Pre-treatment HRF response height in the caudate head and time-to-peak in the caudate tail were top-predictors of treatment response. Time-to-peak in the caudate tail, a region not typically identified in OCD studies using conventional fMRI activation or connectivity measures, may carry novel importance. Additionally, pre-treatment response height in caudate head predicted post-treatment OCD severity (R = -0.48, P = 0.001), and was associated with treatment-related OCD severity changes (R = -0.44, P = 0.0028), underscoring its relevance. With HRF being a reliable marker sensitive to brain function, OCD pathology, and intervention-related changes, these results could guide future studies towards novel discoveries not possible through conventional fMRI approaches like standard BOLD activation or connectivity.

Is there any incremental benefit to conducting neuroimaging and neurocognitive assessments in the diagnosis of ADHD in young children? A machine learning investigation

Given the negative trajectories of early behavior problems associated with ADHD, early diagnosis is considered critical to enable intervention and treatment. To this end, the current investigation employed machine learning to evaluate the relative predictive value of parent/teacher ratings, behavioral and neural measures of executive function (EF) in predicting ADHD in a sample consisting of 162 young children (ages 4–7, mean age 5.55, 82.6 % Hispanic/Latino). Among the target measures, teacher ratings of EF were the most predictive of ADHD. While a more extensive evaluation of neural measures, such as diffusion-weighted imaging, may provide more information as they relate to the underlying cognitive deficits associated with ADHD, the current study indicates that measures of cortical anatomy obtained in research studies, as well cognitive measures of EF often obtained in routine assessments, have little incremental value in differentiating typically developing children from those diagnosed with ADHD. It is important to note that the overlap between some of the EF questions in the BRIEF, and the ADHD symptoms could be enhancing this effect. Thus, future research evaluating the importance of such measures in predicting children’s functional impairment in academic and social areas would provide additional insight into their contributing role in ADHD.

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.

Transcranial photobiomodulation in the management of brain disorders

Transcranial photobiomodulation (tPBM) is the process of delivering light photons through the skull to benefit from its modifying effect. Brain disorders are important health problems. The aim of this review was to determine the existing evidence of effectiveness, useful parameters, and safety of tPBM in the management of traumatic brain injury, stroke, Parkinson, and Alzheimer's disease as the common brain disorders. Four online databases, including Cochrane, Pub Med, Embase, and Google scholar were searched according to the Preferred Reporting Items for Systematic Reviews and meta-analyses (PRISMA) guidelines. 4728 articles were obtained in the initial search. Only those articles that were published until September 2020 and designed as randomized clinical trials (RCTs) or animal-controlled studies were included. 6 RCTs, 2 related supplementary articles, and 38 controlled animal studies met the inclusion criteria of this study. No RCTs were performed in the fields of Alzheimer's and Parkinson's diseases. The human RCTs and animal studies reported no adverse events resulted from the use of tPBM. Useful parameters of tPBM were identified according to the controlled animal studies. Since the investigated RCTs had no homogenous results, making an evidence-based decision for definite therapeutic application of tPBM is still unattainable. Altogether, these data support the need for large confirmatory well-designed RCTs for using tPBM as a novel, safe, and easy-to-administer treatment of brain disorders.

EVIDENCE BEFORE THIS STUDY

High prevalence and complications of brain disorders and also side effects of neuropsychiatric medications have encouraged researchers to find alternative therapeutic techniques which tPBM can be one of them. In present review we tried to determine the existing evidence of effectiveness, useful parameters, and safety of tPBM in the management of traumatic brain injury, stroke, Alzheimer, and Parkinson's disease as common brain disorders. Four online databases, including "Cochrane", "Pub Med", "Embase", and "Google scholar" were searched. Only those articles that were published until September 2020 and designed as RCTs or animal-controlled studies were included. Search keywords were the followings: transcranial photobiomodulation" OR "transcranial low-level laser therapy" AND "stroke" OR "traumatic brain injury" OR "Alzheimer" OR "Parkinson". Several studies have confirmed effectiveness of tPBM in treatment of different brain disorders but the level of evidence of its effectiveness remain to be determined.

ADDED VALUE OF THIS STUDY

In this study we systematically reviewed human RCTs to determine the existing evidence of tPBM effectiveness in management of four mentioned brain disorders. Since the outcomes of the reviewed RCTs were not homogeneous, further well-designed RCTs are required to decide more definitively on the evidence of this noninvasive and probably safe therapeutic intervention. We hypothesized that non-homogeneous outcomes could be due to inefficiency of PBM parameters. Controlled animal studies have the advantage of using objective tests to evaluate the results and compare them with the control group. We determined useful tPBM parameters based on these studies.

IMPLICATIONS OF ALL THE AVAILABLE EVIDENCE

This research is part of our main project of tinnitus treatment using photobiomodulation (PBM). Evidence of central nervous system involvement in tinnitus led us to believe that treatment protocol of tinnitus should also include transcranial PBM. The determined useful parameters can be helpful in designing more efficient tPBM protocols in the management of brain disorders and tinnitus as a common debilitating symptom that can be associated with these disorders.

A unified framework for personalized regions selection and functional relation modeling for early MCI identification

Resting-state functional magnetic resonance imaging (rs-fMRI) has been widely adopted to investigate functional abnormalities in brain diseases. Rs-fMRI data is unsupervised in nature because the psychological and neurological labels are coarse-grained, and no accurate region-wise label is provided along with the complex co-activities of multiple regions. To the best of our knowledge, most studies regarding univariate group analysis or multivariate pattern recognition for brain disease identification have focused on discovering functional characteristics shared across subjects; however, they have paid less attention to individual properties of neural activities that result from different symptoms or degrees of abnormality. In this work, we propose a novel framework that can identify subjects with early-stage mild cognitive impairment (eMCI) and consider individual variability by learning functional relations from automatically selected regions of interest (ROIs) for each subject concurrently. In particular, we devise a deep neural network composed of a temporal embedding module, an ROI selection module, and a disease-identification module. Notably, the ROI selection module is equipped with a reinforcement learning mechanism so it adaptively selects ROIs to facilitate the learning of discriminative feature representations from a temporally embedded blood-oxygen-level-dependent signals. Furthermore, our method allows us to capture the functional relations of a subject-specific ROI subset through the use of a graph-based neural network. Our method considers individual characteristics for diagnosis, as opposed to most conventional methods that identify the same biomarkers across subjects within a group. Based on the ADNI cohort, we validate the effectiveness of our method by presenting the superior performance of our network in eMCI identification. Furthermore, we provide insightful neuroscientific interpretations by analyzing the regions selected for the eMCI classification.

A Deep Learning Approach to Predict Autism Spectrum Disorder Using Multisite Resting-State fMRI

Autism spectrum disorder (ASD) is a complex and degenerative neuro-developmental disorder. Most of the existing methods utilize functional magnetic resonance imaging (fMRI) to detect ASD with a very limited dataset which provides high accuracy but results in poor generalization. To overcome this limitation and to enhance the performance of the automated autism diagnosis model, in this paper, we propose an ASD detection model using functional connectivity features of resting-state fMRI data. Our proposed model utilizes two commonly used brain atlases, Craddock 200 (CC200) and Automated Anatomical Labelling (AAL), and two rarely used atlases Bootstrap Analysis of Stable Clusters (BASC) and Power. A deep neural network (DNN) classifier is used to perform the classification task. Simulation results indicate that the proposed model outperforms state-of-the-art methods in terms of accuracy. The mean accuracy of the proposed model was 88%, whereas the mean accuracy of the state-of-the-art methods ranged from 67% to 85%. The sensitivity, F1-score, and area under receiver operating characteristic curve (AUC) score of the proposed model were 90%, 87%, and 96%, respectively. Comparative analysis on various scoring strategies show the superiority of BASC atlas over other aforementioned atlases in classifying ASD and control.

Resample aggregating improves the generalizability of connectome predictive modeling

It is a longstanding goal of neuroimaging to produce reliable, generalizable models of brain behavior relationships. More recently, data driven predictive models have become popular. However, overfitting is a common problem with statistical models, which impedes model generalization. Cross validation (CV) is often used to estimate expected model performance within sample. Yet, the best way to generate brain behavior models, and apply them out-of-sample, on an unseen dataset, is unclear. As a solution, this study proposes an ensemble learning method, in this case resample aggregating, encompassing both model parameter estimation and feature selection. Here we investigate the use of resampled aggregated models when used to estimate fluid intelligence (fIQ) from fMRI based functional connectivity (FC) data. We take advantage of two large openly available datasets, the Human Connectome Project (HCP), and the Philadelphia Neurodevelopmental Cohort (PNC). We generate aggregated and non-aggregated models of fIQ in the HCP, using the Connectome Prediction Modelling (CPM) framework. Over various test-train splits, these models are evaluated in sample, on left-out HCP data, and out-of-sample, on PNC data. We find that a resample aggregated model performs best both within- and out-of-sample. We also find that feature selection can vary substantially within-sample. More robust feature selection methods, as detailed here, are needed to improve cross sample performance of CPM based brain behavior models.

Psychiatric Neural Networks and Precision Therapeutics by Machine Learning

Learning and environmental adaptation increase the likelihood of survival and improve the quality of life. However, it is often difficult to judge optimal behaviors in real life due to highly complex social dynamics and environment. Consequentially, many different brain regions and neuronal circuits are involved in decision-making. Many neurobiological studies on decision-making show that behaviors are chosen through coordination among multiple neural network systems, each implementing a distinct set of computational algorithms. Although these processes are commonly abnormal in neurological and psychiatric disorders, the underlying causes remain incompletely elucidated. Machine learning approaches with multidimensional data sets have the potential to not only pathologically redefine mental illnesses but also better improve therapeutic outcomes than DSM/ICD diagnoses. Furthermore, measurable endophenotypes could allow for early disease detection, prognosis, and optimal treatment regime for individuals. In this review, decision-making in real life and psychiatric disorders and the applications of machine learning in brain imaging studies on psychiatric disorders are summarized, and considerations for the future clinical translation are outlined. This review also aims to introduce clinicians, scientists, and engineers to the opportunities and challenges in bringing artificial intelligence into psychiatric practice.

Structural MRI and functional connectivity features predict current clinical status and persistence behavior in prescription opioid users

Prescription opioid use disorder (POUD) has reached epidemic proportions in the United States, raising an urgent need for diagnostic biological tools that can improve predictions of disease characteristics. The use of neuroimaging methods to develop such biomarkers have yielded promising results when applied to neurodegenerative and psychiatric disorders, yet have not been extended to prescription opioid addiction. With this long-term goal in mind, we conducted a preliminary study in this understudied clinical group. Univariate and multivariate approaches to distinguishing between POUD (n = 26) and healthy controls (n = 21) were investigated, on the basis of structural MRI (sMRI) and resting-state functional connectivity (restFC) features. Univariate approaches revealed reduced structural integrity in the subcortical extent of a previously reported addiction-related network in POUD subjects. No reliable univariate between-group differences in cortical structure or edgewise restFC were observed. Contrasting these mixed univariate results, multivariate machine learning classification approaches recovered more statistically reliable group differences, especially when sMRI and restFC features were combined in a multi-modal model (classification accuracy = 66.7%, p < .001). The same multivariate multi-modal approach also yielded reliable prediction of individual differences in a clinically relevant behavioral measure (persistence behavior; predicted-to-actual overlap r = 0.42, p = .009). Our findings suggest that sMRI and restFC measures can be used to reliably distinguish the neural effects of long-term opioid use, and that this endeavor numerically benefits from multivariate predictive approaches and multi-modal feature sets. This can serve as theoretical proof-of-concept for future longitudinal modeling of prognostic POUD characteristics from neuroimaging features, which would have clearer clinical utility.

NEURO-LEARN: a Solution for Collaborative Pattern Analysis of Neuroimaging Data

The development of neuroimaging instrumentation has boosted neuroscience researches. Consequently, both the fineness and the cost of data acquisition have profoundly increased, leading to the main bottleneck of this field: limited sample size and high dimensionality of neuroimaging data. Therefore, the emphasis of ideas of data pooling and research collaboration has increased over the past decade. Collaborative analysis techniques emerge as the idea developed. In this paper, we present NEURO-LEARN, a solution for collaborative pattern analysis of neuroimaging data. Its collaboration scheme consists of four parts: projects, data, analysis, and reports. While data preparation workflows defined in projects reduce the high dimensionality of neuroimaging data by collaborative computation, pooling of derived data and sharing of pattern analysis workflows along with generated reports on the Web enlarge the sample size and ensure the reliability and reproducibility of pattern analysis. Incorporating this scheme, NEURO-LEARN provides an easy-to-use Web application that allows users from different sites to share projects and processed data, perform pattern analysis, and obtain result reports. We anticipate that this solution will help neuroscientists to enlarge sample size, conquer the curse of dimensionality and conduct reproducible studies on neuroimaging data with efficiency and validity.

A Mutual Multi-Scale Triplet Graph Convolutional Network for Classification of Brain Disorders Using Functional or Structural Connectivity

Brain connectivity alterations associated with mental disorders have been widely reported in both functional MRI (fMRI) and diffusion MRI (dMRI). However, extracting useful information from the vast amount of information afforded by brain networks remains a great challenge. Capturing network topology, graph convolutional networks (GCNs) have demonstrated to be superior in learning network representations tailored for identifying specific brain disorders. Existing graph construction techniques generally rely on a specific brain parcellation to define regions-of-interest (ROIs) to construct networks, often limiting the analysis into a single spatial scale. In addition, most methods focus on the pairwise relationships between the ROIs and ignore high-order associations between subjects. In this letter, we propose a mutual multi-scale triplet graph convolutional network (MMTGCN) to analyze functional and structural connectivity for brain disorder diagnosis. We first employ several templates with different scales of ROI parcellation to construct coarse-to-fine brain connectivity networks for each subject. Then, a triplet GCN (TGCN) module is developed to learn functional/structural representations of brain connectivity networks at each scale, with the triplet relationship among subjects explicitly incorporated into the learning process. Finally, we propose a template mutual learning strategy to train different scale TGCNs collaboratively for disease classification. Experimental results on 1,160 subjects from three datasets with fMRI or dMRI data demonstrate that our MMTGCN outperforms several state-of-the-art methods in identifying three types of brain disorders.

Hippocampal and Amygdalar Morphological Abnormalities in Alzheimer's Disease Based on Three Chinese MRI Datasets

BACKGROUND

Early detection of Alzheimer's disease (AD) and its early stage, the mild cognitive impairment (MCI), has important scientific, clinical and social significance. Magnetic resonance imaging (MRI) based statistical shape analysis provides an opportunity to detect regional structural abnormalities of brain structures caused by AD and MCI.

OBJECTIVE

In this work, we aimed to employ a well-established statistical shape analysis pipeline, in the framework of large deformation diffeomorphic metric mapping, to identify and quantify the regional shape abnormalities of the bilateral hippocampus and amygdala at different prodromal stages of AD, using three Chinese MRI datasets collected from different domestic hospitals.

METHODS

We analyzed the region-specific shape abnormalities at different stages of the neuropathology of AD by comparing the localized shape characteristics of the bilateral hippocampi and amygdalas between healthy controls and two disease groups (MCI and AD). In addition to group comparison analyses, we also investigated the association between the shape characteristics and the Mini Mental State Examination (MMSE) of each structure of interest in the disease group (MCI and AD combined) as well as the discriminative power of different morphometric biomarkers.

RESULTS

We found the strongest disease pathology (regional atrophy) at the subiculum and CA1 subregions of the hippocampus and the basolateral, basomedial as well as centromedial subregions of the amygdala. Furthermore, the shape characteristics of the hippocampal and amygdalar subregions exhibiting the strongest AD related atrophy were found to have the most significant positive associations with the MMSE. Employing the shape deformation marker of the hippocampus or the amygdala for automated MCI or AD detection yielded a significant accuracy boost over the corresponding volume measurement.

CONCLUSION

Our results suggested that the amygdalar and hippocampal morphometrics, especially those of shape morphometrics, can be used as auxiliary indicators for monitoring the disease status of an AD patient.

Machine Learning Analytics of Resting-State Functional Connectivity Predicts Survival Outcomes of Glioblastoma Multiforme Patients

Glioblastoma multiforme (GBM) is the most frequently occurring brain malignancy. Due to its poor prognosis with currently available treatments, there is a pressing need for easily accessible, non-invasive techniques to help inform pre-treatment planning, patient counseling, and improve outcomes. In this study we determined the feasibility of resting-state functional connectivity (rsFC) to classify GBM patients into short-term and long-term survival groups with respect to reported median survival (14.6 months). We used a support vector machine with rsFC between regions of interest as predictive features. We employed a novel hybrid feature selection method whereby features were first filtered using correlations between rsFC and OS, and then using the established method of recursive feature elimination (RFE) to select the optimal feature subset. Leave-one-subject-out cross-validation evaluated the performance of models. Classification between short- and long-term survival accuracy was 71.9%. Sensitivity and specificity were 77.1 and 65.5%, respectively. The area under the receiver operating characteristic curve was 0.752 (95% CI, 0.62–0.88). These findings suggest that highly specific features of rsFC may predict GBM survival. Taken together, the findings of this study support that resting-state fMRI and machine learning analytics could enable a radiomic biomarker for GBM, augmenting care and planning for individual patients.

Deep learning applications for the classification of psychiatric disorders using neuroimaging data: Systematic review and meta-analysis

Deep learning (DL) methods have been increasingly applied to neuroimaging data to identify patients with psychiatric and neurological disorders. This review provides an overview of the different DL applications within psychiatry and compares DL model accuracy to standard machine learning (SML). Fifty-three articles were included for qualitative analysis, primarily investigating autism spectrum disorder (ASD; n = 22), schizophrenia (SZ; n = 22) and attention-deficit/hyperactivity disorder (ADHD; n = 9). Thirty-two of the thirty-five studies that directly compared DL to SML reported a higher accuracy for DL. Only sixteen studies could be included in a meta-regression to quantitatively compare DL and SML performance. This showed a higher odds ratio for DL models, though the comparison attained significance only for ASD. Our results suggest that deep learning of neuroimaging data is a promising tool for the classification of individual psychiatric patients. However, it is not yet used to its full potential: most studies use pre-engineered features, whereas one of the main advantages of DL is its ability to learn representations of minimally processed data. Our current evaluation is limited by minimal reporting of performance measures to enable quantitative comparisons, and the restriction to ADHD, SZ and ASD as current research focusses on large publicly available datasets. To truly uncover the added value of DL, we need carefully designed comparisons of SML and DL models which are yet rarely performed.

Brain dynamics of (a)typical reading development-a review of longitudinal studies

Literacy development is a process rather than a single event and thus should be studied at multiple time points. A longitudinal design employing neuroimaging methods offers the possibility to identify neural changes associated with reading development, and to reveal early markers of dyslexia. The core of this review is a summary of findings from longitudinal neuroimaging studies on typical and atypical reading development. Studies focused on the prediction of reading gains with a single neuroimaging time point complement this review. Evidence from structural studies suggests that reading development results in increased structural integrity and functional specialization of left-hemispheric language areas. Compromised integrity of some of these tracts in children at risk for dyslexia might be compensated by higher anatomical connectivity in the homologous right hemisphere tracts. Regarding function, activation in phonological and audiovisual integration areas and growing sensitivity to print in the ventral occipito-temporal cortex (vOT) seem to be relevant neurodevelopmental markers of successful reading acquisition. Atypical vOT responses at the beginning of reading training and infant auditory brain potentials have been proposed as neuroimaging predictors of dyslexia that can complement behavioral measures. Besides these insights, longitudinal neuroimaging studies on reading and dyslexia are still relatively scarce and small sample sizes raise legitimate concerns about the reliability of the results. This review discusses the challenges of these studies and provides recommendations to improve this research area. Future longitudinal research with larger sample sizes are needed to improve our knowledge of typical and atypical reading neurodevelopment.

Abnormal functional connectivity within the reward network: a potential neuroimaging endophenotype of bipolar disorder

BACKGROUND

Reward circuit dysfunction underlies the pathogenesis of bipolar disorder (BD). This study aims to investigate whether nucleus accumbens (NAcc) and ventromedial prefrontal cortex (vmPFC), two key reward regions for BD, have resting-state dysfunctional connectivity with other brain regions in depressed and euthymic BD.

METHODS

40 bipolar depressive (DE), 20 euthymic patients (EU) and 20 healthy controls (HC) were recruited to undergo resting-state functional MRI (rs-fMRI) scanning. Seed-based functional connectivity (FC) was calculated between NAcc/vmPFC and the whole brain. Group differences were calculated and their correlations with clinical characteristics were analyzed. Support vector machine was applied to classify BD patients and HC based on the FC between the cluster of group difference and NAcc/vmPFC.

RESULTS

Whole brain networks of FC identified right anterior insular cortex (AIC) as a significant region with bilateral NAcc when compared among three groups. The right AIC-NAcc FC elevated in both patient groups and was highest in the EU group. Interestingly, vmPFC-based networks also identified the right AIC as a significant cluster. The right AIC-vmPFC FC elevated in both patient groups. However, FC between NAcc and vmPFC did not significantly differ BD patients from HC. Furthermore, the strength of FC between bilateral NAcc and the right AIC was positively associated with the illness course of BD. Notably, the NAcc/vmPFC-right AIC classifier acquired an accuracy of 68.75% and AUC-ROC of 78.17%.

LIMITATIONS

Our sample size is modest.

CONCLUSIONS

Our findings indicated that elevated NAcc/vmPFC-right AIC connectivity within the reward circuit could be a neuroimaging endophenotype of BD.

Promises and pitfalls of deep neural networks in neuroimaging-based psychiatric research

By promising more accurate diagnostics and individual treatment recommendations, deep neural networks and in particular convolutional neural networks have advanced to a powerful tool in medical imaging. Here, we first give an introduction into methodological key concepts and resulting methodological promises including representation and transfer learning, as well as modelling domain-specific priors. After reviewing recent applications within neuroimaging-based psychiatric research, such as the diagnosis of psychiatric diseases, delineation of disease subtypes, normative modeling, and the development of neuroimaging biomarkers, we discuss current challenges. This includes for example the difficulty of training models on small, heterogeneous and biased data sets, the lack of validity of clinical labels, algorithmic bias, and the influence of confounding variables.

Cognitive and MRI trajectories for prediction of Alzheimer's disease

The concept of Mild Cognitive Impairment (MCI) is used to describe the early stages of Alzheimer's disease (AD), and identification and treatment before further decline is an important clinical task. We selected longitudinal data from the ADNI database to investigate how well normal function (HC, n= 134) vs. conversion to MCI (cMCI, n= 134) and stable MCI (sMCI, n=333) vs. conversion to AD (cAD, n= 333) could be predicted from cognitive tests, and whether the predictions improve by adding information from magnetic resonance imaging (MRI) examinations. Features representing trajectories of change in the selected cognitive and MRI measures were derived from mixed effects models and used to train ensemble machine learning models to classify the pairs of subgroups based on a subset of the data set. Evaluation in an independent test set showed that the predictions for HC vs. cMCI improved substantially when MRI features were added, with an increase in [Formula: see text]-score from 60 to 77%. The [Formula: see text]-scores for sMCI vs. cAD were 77% without and 78% with inclusion of MRI features. The results are in-line with findings showing that cognitive changes tend to manifest themselves several years after the Alzheimer's disease is well-established in the brain.

Neural reference groups: a synchrony-based classification approach for predicting attitudes using fNIRS

Social neuroscience research has demonstrated that those who are like-minded are also 'like-brained.' Studies have shown that people who share similar viewpoints have greater neural synchrony with one another, and less synchrony with people who 'see things differently.' Although these effects have been demonstrated at the 'group level,' little work has been done to predict the viewpoints of specific 'individuals' using neural synchrony measures. Furthermore, the studies that have made predictions using synchrony-based classification at the individual level used expensive and immobile neuroimaging equipment (e.g. functional magnetic resonance imaging) in highly controlled laboratory settings, which may not generalize to real-world contexts. Thus, this study uses a simple synchrony-based classification method, which we refer to as the 'neural reference groups' approach, to predict individuals' dispositional attitudes from data collected in a mobile 'pop-up neuroscience' lab. Using functional near-infrared spectroscopy data, we predicted individuals' partisan stances on a sociopolitical issue by comparing their neural timecourses to data from two partisan neural reference groups. We found that partisan stance could be identified at above-chance levels using data from dorsomedial prefrontal cortex. These results indicate that the neural reference groups approach can be used to investigate naturally occurring, dispositional differences anywhere in the world.

Deep learning encodes robust discriminative neuroimaging representations to outperform standard machine learning

Recent critical commentaries unfavorably compare deep learning (DL) with standard machine learning (SML) approaches for brain imaging data analysis. However, their conclusions are often based on pre-engineered features depriving DL of its main advantage — representation learning. We conduct a large-scale systematic comparison profiled in multiple classification and regression tasks on structural MRI images and show the importance of representation learning for DL. Results show that if trained following prevalent DL practices, DL methods have the potential to scale particularly well and substantially improve compared to SML methods, while also presenting a lower asymptotic complexity in relative computational time, despite being more complex. We also demonstrate that DL embeddings span comprehensible task-specific projection spectra and that DL consistently localizes task-discriminative brain biomarkers. Our findings highlight the presence of nonlinearities in neuroimaging data that DL can exploit to generate superior task-discriminative representations for characterizing the human brain.

Recent critical commentaries unfavorably compare deep learning (DL) with standard machine learning (SML) for brain imaging data analysis. Here, the authors show that if trained following prevalent DL practices, DL methods substantially improve compared to SML methods by encoding robust discriminative brain representations.

From Personalized Medicine to Precision Psychiatry?

Personalised medicine aims to find an individualized approach for each particular patient. Most factors used in current psychiatry, however, depend on the assessment made by the individual clinician and lack a higher degree of reliability. Precision medicine bases decisions on quantifiable indicators available thanks to the tremendous progress in science and technology facilitating the acquisition, processing and analysis of huge amounts of data. So far, psychiatry has not been benefiting enough from the advanced diagnostic technologies; nevertheless, we are witnessing the dawn of the era of precision psychiatry, starting with the gathering of sufficient amounts of data and its analysis by the means of artificial intelligence and machine learning. First results of this approach in psychiatry are available, which facilitate diagnosis assessment, course prediction, and appropriate treatment choice. These processes are often so complex and difficult to understand that they may resemble a "black box", which can slow down the acceptance of the results of this approach in clinical practice. Still, bringing precision medicine including psychiatry to standard clinical practice is a big challenge that can result in a completely new and transformative concept of health care. Such extensive changes naturally have both their supporters and opponents. This paper aims to familiarize clinically oriented physicians with precision psychiatry and to attract their attention to its recent developments. We cover the theoretical basis of precision medicine, its specifics in psychiatry, and provide examples of its use in the field of diagnostic assessment, course prediction, and appropriate treatment planning.

Deep Neural Network to Differentiate Brain Activity Between Patients With First-Episode Schizophrenia and Healthy Individuals: A Multi-Channel Near Infrared Spectroscopy Study

Backgrounds: Reduced brain cortical activity over the frontotemporal regions measured by near infrared spectroscopy (NIRS) has been reported in patients with first-episode schizophrenia (FES). This study aimed to differentiate between patients with FES and healthy controls (HCs) on basis of the frontotemporal activity measured by NIRS with a support vector machine (SVM) and deep neural network (DNN) classifier. In addition, we compared the accuracy of performance of SVM and DNN. Methods: In total, 33 FES patients and 34 HCs were recruited. Their brain cortical activities were measured using NIRS while performing letter and category versions of verbal fluency tests (VFTs). The integral and centroid values of brain cortical activity in the bilateral frontotemporal regions during the VFTs were selected as features in SVM and DNN classifier. Results: Compared to HCs, FES patients displayed reduced brain cortical activity over the bilateral frontotemporal regions during both types of VFTs. Regarding the classifier performance, SVM reached an accuracy of 68.6%, sensitivity of 70.1%, and specificity of 64.6%, while DNN reached an accuracy of 79.7%, sensitivity of 88.8%, and specificity of 74.9% in the classification of FES patients and HCs. Conclusions: Compared to findings of previous structural neuroimaging studies, we found that using DNN to measure the NIRS signals during the VFTs to differentiate between FES patients and HCs could achieve a higher accuracy, indicating that NIRS can be used as a potential marker to classify FES patients from HCs. Future additional independent datasets are needed to confirm the validity of our model.

Machine learning for psychiatry: getting doctors at the black box?

Recent developments in the field of machine learning have spurred high hopes for diagnostic support for psychiatric patients based on brain MRI. But while technical advances are undoubtedly remarkable, the current trajectory of mostly proof-of-concept studies performed on retrospective, often repository-derived data, may not be well suited to yield a substantial impact in clinical practice. Here we review these developments and challenges, arguing for the need of stronger involvement of and input from medical doctors in order to pave the way for machine learning in clinical psychiatry.

Blinded Clinical Evaluation for Dementia of Alzheimer's Type Classification Using FDG-PET: A Comparison Between Feature-Engineered and Non-Feature-Engineered Machine Learning Methods

BACKGROUND

Advanced machine learning methods can aid in the identification of dementia risk using neuroimaging-derived features including FDG-PET. However, to enable the translation of these methods and test their usefulness in clinical practice, it is crucial to conduct independent validation on real clinical samples, which has yet to be properly delineated in the current literature.

OBJECTIVE

In this paper, we present our efforts to enable such clinical translational through the evaluation and comparison of two machine-learning methods for discrimination between dementia of Alzheimer's type (DAT) and Non-DAT controls.

METHODS

FDG-PET-based dementia scores were generated on an independent clinical sample whose clinical diagnosis was blinded to the algorithm designers. A feature-engineered approach (multi-kernel probability classifier) and a non-feature-engineered approach (3D convolutional neural network) were analyzed. Both classifiers were pre-trained on cognitively normal subjects as well as subjects with DAT. These two methods provided a probabilistic dementia score for this previously unseen clinical data. Performance of the algorithms were compared against ground-truth dementia rating assessed by experienced nuclear physicians.

RESULTS

Blinded clinical evaluation on both classifiers showed good separation between the cognitively normal subjects and the patients diagnosed with DAT. The non-feature-engineered dementia score showed higher sensitivity among subjects whose diagnosis was in agreement between the machine-learning models, while the feature-engineered approach showed higher specificity in non-consensus cases.

CONCLUSION

In this study, we demonstrated blinded evaluation using data from an independent clinical sample for assessing the performance in DAT classification models in a clinical setting. Our results showed good generalizability for two machine-learning approaches, marking an important step for the translation of pre-trained machine-learning models into clinical practice.

Dynamic functional network reconfiguration underlying the pathophysiology of schizophrenia and autism spectrum disorder

The dynamics of the human brain span multiple spatial scales, from connectivity associated with a specific region/network to the global organization, each representing different brain mechanisms. Yet brain reconfigurations at different spatial scales are seldom explored and whether they are associated with the neural aspects of brain disorders is far from understood. In this study, we introduced a dynamic measure called step-wise functional network reconfiguration (sFNR) to characterize how brain configuration rewires at different spatial scales. We applied sFNR to two independent datasets, one includes 160 healthy controls (HCs) and 151 patients with schizophrenia (SZ) and the other one includes 314 HCs and 255 individuals with autism spectrum disorder (ASD). We found that both SZ and ASD have increased whole-brain sFNR and sFNR between cerebellar and subcortical/sensorimotor domains. At the ICN level, the abnormalities in SZ are mainly located in ICNs within subcortical, sensory, and cerebellar domains, while the abnormalities in ASD are more widespread across domains. Interestingly, the overlap SZ-ASD abnormality in sFNR between cerebellar and sensorimotor domains was correlated with the reasoning-problem-solving performance in SZ (r = -.1652, p = .0058) as well as the Autism Diagnostic Observation Schedule in ASD (r = .1853, p = .0077). Our findings suggest that dynamic reconfiguration deficits may represent a key intersecting point for SZ and ASD. The investigation of brain dynamics at different spatial scales can provide comprehensive insights into the functional reconfiguration, which might advance our knowledge of cognitive decline and other pathophysiology in brain disorders.

Multi-scale graph-based grading for Alzheimer's disease prediction

The prediction of subjects with mild cognitive impairment (MCI) who will progress to Alzheimer's disease (AD) is clinically relevant, and may above all have a significant impact on accelerating the development of new treatments. In this paper, we present a new MRI-based biomarker that enables us to accurately predict conversion of MCI subjects to AD. In order to better capture the AD signature, we introduce two main contributions. First, we present a new graph-based grading framework to combine inter-subject similarity features and intra-subject variability features. This framework involves patch-based grading of anatomical structures and graph-based modeling of structure alteration relationships. Second, we propose an innovative multiscale brain analysis to capture alterations caused by AD at different anatomical levels. Based on a cascade of classifiers, this multiscale approach enables the analysis of alterations of whole brain structures and hippocampus subfields at the same time. During our experiments using the ADNI-1 dataset, the proposed multiscale graph-based grading method obtained an area under the curve (AUC) of 81% to predict conversion of MCI subjects to AD within three years. Moreover, when combined with cognitive scores, the proposed method obtained 85% of AUC. These results are competitive in comparison to state-of-the-art methods evaluated on the same dataset.

Neuroimaging-based brain-age prediction in diverse forms of epilepsy: a signature of psychosis and beyond

Epilepsy is a diverse brain disorder, and the pathophysiology of its various forms and comorbidities is largely unknown. A recent machine learning method enables us to estimate an individual's "brain-age" from MRI; this brain-age prediction is expected as a novel individual biomarker of neuropsychiatric disorders. The aims of this study were to estimate the brain-age for various categories of epilepsy and to evaluate clinical discrimination by brain-age for (1) the effect of psychosis on temporal lobe epilepsy (TLE), (2) psychogenic nonepileptic seizures (PNESs) from MRI-negative epilepsies, and (3) progressive myoclonic epilepsy (PME) from juvenile myoclonic epilepsy (JME). In total, 1196 T1-weighted MRI scans from healthy controls (HCs) were used to build a brain-age prediction model with support vector regression. Using the model, we calculated the brain-predicted age difference (brain-PAD: predicted age-chronological age) of the HCs and 318 patients with epilepsy. We compared the brain-PAD values based on the research questions. As a result, all categories of patients except for extra-temporal lobe focal epilepsy showed a significant increase in brain-PAD. TLE with hippocampal sclerosis presented a significantly higher brain-PAD than several other categories. The mean brain-PAD in TLE with inter-ictal psychosis was 10.9 years, which was significantly higher than TLE without psychosis (5.3 years). PNES showed a comparable mean brain-PAD (10.6 years) to that of epilepsy patients. PME had a higher brain-PAD than JME (22.0 vs. 9.3 years). In conclusion, neuroimaging-based brain-age prediction can provide novel insight into or clinical usefulness for the diverse symptoms of epilepsy.

Graph convolution network with similarity awareness and adaptive calibration for disease-induced deterioration prediction

Graph convolution networks (GCN) have been successfully applied in disease prediction tasks as they capture interactions (i.e., edges and edge weights on the graph) between individual elements. The interactions in existing works are constructed by fusing similarity between imaging information and distance between non-imaging information, whereas disregarding the disease status of those individuals in the training set. Besides, the similarity is being evaluated by computing the correlation distance between feature vectors, which limits prediction performance, especially for predicting significant memory concern (SMC) and mild cognitive impairment (MCI). In this paper, we propose three mechanisms to improve GCN, namely similarity-aware adaptive calibrated GCN (SAC-GCN), for predicting SMC and MCI. First, we design a similarity-aware graph using different receptive fields to consider disease status. The labelled subjects on the graph are only connected with those labelled subjects with the same status. Second, we propose an adaptive mechanism to evaluate similarity. Specifically, we construct initial GCN with evaluating similarity by using traditional correlation distance, then pre-train the initial GCN by using training samples and use it to score all subjects. Then, the difference between these scores replaces correlation distance to update similarity. Last, we devise a calibration mechanism to fuse functional magnetic resonance imaging (fMRI) and diffusion tensor imaging (DTI) information into edges. The proposed method is tested on the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset. Experimental results demonstrate that our proposed method is useful to predict disease-induced deterioration and superior to other related algorithms, with a mean classification accuracy of 86.83% in our prediction tasks.

Multimodal MRI assessment for first episode psychosis: A major change in the thalamus and an efficient stratification of a subgroup

Multi‐institutional brain imaging studies have emerged to resolve conflicting results among individual studies. However, adjusting multiple variables at the technical and cohort levels is challenging. Therefore, it is important to explore approaches that provide meaningful results from relatively small samples at institutional levels. We studied 87 first episode psychosis (FEP) patients and 62 healthy subjects by combining supervised integrated factor analysis (SIFA) with a novel pipeline for automated structure‐based analysis, an efficient and comprehensive method for dimensional data reduction that our group recently established. We integrated multiple MRI features (volume, DTI indices, resting state fMRI—rsfMRI) in the whole brain of each participant in an unbiased manner. The automated structure‐based analysis showed widespread DTI abnormalities in FEP and rs‐fMRI differences between FEP and healthy subjects mostly centered in thalamus. The combination of multiple modalities with SIFA was more efficient than the use of single modalities to stratify a subgroup of FEP (individuals with schizophrenia or schizoaffective disorder) that had more robust deficits from the overall FEP group. The information from multiple MRI modalities and analytical methods highlighted the thalamus as significantly abnormal in FEP. This study serves as a proof‐of‐concept for the potential of this methodology to reveal disease underpins and to stratify populations into more homogeneous sub‐groups.

Präzisionspsychiatrie und der Beitrag von Brain Imaging und anderen Biomarkern

'Precision Psychiatry' as the psychiatric variant of 'Precision Medicine' aims to provide high-level diagnosis and treatment based on robust biomarkers and tailored to the individual clinical, neurobiological, and genetic constitution of the patient. The specific peculiarity of psychiatry, in which disease entities are normatively defined based on clinical experience and are also significantly influenced by contemporary history, society and philosophy, has so far made the search for valid and reliable psychobiological connections difficult. Nevertheless, considerable progress has now been made in all areas of psychiatric research, made possible above all by the critical review and renewal of previous concepts of disease and psychopathology, the increased orientation towards neurobiology and genetics, and in particular the use of machine learning methods. Notably, modern machine learning methods make it possible to integrate high-dimensional and multimodal data sets and generate models which provide new psychobiological insights and offer the possibility of individualized, biomarker-driven single-subject prediction of diagnosis, therapy response and prognosis. The aim of the present review is therefore to introduce the concept of 'Precision Psychiatry' to the interested reader, to concisely present modern, machine learning methods required for this, and to clearly present the current state and future of biomarker-based 'precision psychiatry'.