Prediction of activation patterns preceding hallucinations in patients with schizophrenia using machine learning with structured sparsity

Understanding visual hallucinations: a new synthesis

Despite decades of research, we do not definitively know how people sometimes see things that are not there. Eight models of complex visual hallucinations have been published since 2000, including Deafferentation, Reality Monitoring, Perception and Attention Deficit, Activation, Input, and Modulation, Hodological, Attentional Networks, Active inference, and Thalamocortical Dysrhythmia Default Mode Network Decoupling. Each was derived from different understandings of brain organisation. To reduce this variability, representatives from each research group agreed an integrated Visual Hallucination Framework that is consistent with current theories of veridical and hallucinatory vision. The Framework delineates cognitive systems relevant to hallucinations. It allows a systematic, consistent, investigation of relationships between the phenomenology of visual hallucinations and changes in underpinning cognitive structures. The episodic nature of hallucinations highlights separate factors associated with the onset, persistence, and end of specific hallucinations suggesting a complex relationship between state and trait markers of hallucination risk. In addition to a harmonised interpretation of existing evidence, the Framework highlights new avenues of research, and potentially, new approaches to treating distressing hallucinations.

Time varying dynamics of hallucinations in clinical and non-clinical voice-hearers

Auditory verbal hallucinations (AVH) are frequently associated with psychotic disorders, yet also occur in non-clinical voice-hearers. AVH in this group are similar to those within clinical voice-hearers in terms of several phenomenological aspects, but non-clinical voice-hearers report to have more control over their AVH and attribute less emotional valence to them. These dissimilarities may stem from differences on the neurobiological level, as it is still under debate whether the mechanisms involved in AVH are the same in clinical and non-clinical voice-hearers. In this study, 21 clinical and 21 non-clinical voice-hearers indicated the onset and offsets of AVH during an fMRI scan. Using a method called leading eigenvector dynamics analysis (LEiDA), we examined time-varying dynamics of functional connectivity involved in AVH with a sub-second temporal resolution. We assessed differences between groups, and between hallucination and rest periods in dwell time, switching frequency, probability of occurrence, and transition probabilities of nine recurrent states of functional connectivity with a permutation ANOVA. Deviations in dwell times, switching frequencies, and switch probabilities in the hallucination period indicated more erratic dynamics during this condition regardless of their clinical status. Post-hoc analyses of the dwell times exhibited the most distinct differences between the rest and hallucination condition for the non-clinical sample, suggesting stronger differences between the two conditions in this group. Overall, these findings suggest that the neurobiological mechanisms involved in AVH are similar in clinical and non-clinical individuals.

Feasibility and usefulness of brain imaging in catatonia

Catatonia is a well characterized psychomotor syndrome that has recognizable motor, affective, behavioural and vegetative manifestations. Despite recent demonstration that catatonia is often associated with brain imaging abnormalities, there is currently no consensus or guidelines about the role of brain imaging. In this study, we assessed the feasibility of brain imaging in a series of patients with catatonia in a routine clinical setting and estimated the prevalence of clinically relevant radiological abnormalities. Sixty patients with catatonia were evaluated against sixty non-healthy controls subjects with headache. The MRI reports were reviewed, and MRI scans were also interpreted by neuroradiologists using a standardised MRI assessment. In this cohort, more than 85% of brain scans of patients with catatonia revealed abnormalities. The most frequently reported abnormalities in the catatonic group were white matter abnormalities (n = 44), followed by brain atrophy (n = 27). There was no evidence for significant differences in the frequency of abnormalities found in radiology reports and standardised neuroradiological assessments. The frequency of abnormalities was similar to that found in a population of non-healthy controls subjects with headache. This study shows that MRI is feasible in patients with catatonia and that brain imaging abnormalities are common findings in these patients. Most frequently, white matter abnormalities and diffuse brain atrophy are observed.

Neural Activation in the Ventromedial Prefrontal Cortex Precedes Conscious Experience of Being in or out of a Transient Hallucinatory State

BACKGROUND AND HYPOTHESES

Auditory verbal hallucinations (AVHs) is not only a common symptom in schizophrenia but also observed in individuals in the general population. Despite extensive research, AVHs are poorly understood, especially their underlying neuronal architecture. Neuroimaging methods have been used to identify brain areas and networks that are activated during hallucinations. A characteristic feature of AVHs is, however, that they fluctuate over time, with varying frequencies of starts and stops. An unanswered question is, therefore, what neuronal events co-occur with the initiation and inhibition of an AVH episode.

STUDY DESIGN

We investigated brain activation with fMRI in 66 individuals who experienced multiple AVH-episodes while in the scanner. We extracted time-series fMRI-data and monitored changes second-by-second from 10 s before to 15 s after participants indicated the start and stop of an episode, respectively, by pressing a hand-held response-button.

STUDY RESULTS

We found a region in the ventromedial prefrontal cortex (VMPFC) which showed a significant increase in activation initiated a few seconds before participants indicated the start of an episode, and a corresponding decrease in activation initiated a few seconds before the end of an episode.

CONCLUSIONS

The consistent increase and decrease in activation in this area in advance of the consciously experienced presence or absence of the "voice" imply that this region may act as a switch in turning episodes on and off. The activation is unlikely to be confounded by motor responses. The findings could have clinical implications for brain stimulation treatments, like transcranial magnetic stimulation.

Intrusive experiences in posttraumatic stress disorder: Treatment response induces changes in the directed functional connectivity of the anterior insula

•

Many causal paths were less influenced by the AI after effective therapy for PTSD.

•

Insular influences over the rest of the brain were found to be positively correlated with re-experiencing.

•

Re-experiencing was linked with changes in intrinsic networks’ spatial stability after treatment.

Background

One of the core features of posttraumatic stress disorder (PTSD) is re-experiencing trauma. The anterior insula (AI) has been proposed to play a crucial role in these intrusive experiences. However, the dynamic function of the AI in re-experiencing trauma and its putative modulation by effective therapy need to be specified.

Methods

Thirty PTSD patients were enrolled and exposed to traumatic memory reactivation therapy. Resting-state functional magnetic resonance imaging (fMRI) scans were acquired before and after treatment. To explore AI-directed influences over the rest of the brain, we referred to a mixed model using pre-/posttreatment Granger causality analysis seeded on the AI as a within-subject factor and treatment response as a between-subject factor. To further identify correlates of re-experiencing trauma, we investigated how intrusive severity affected (i) causality maps and (ii) the spatial stability of other intrinsic brain networks.

Results

We observed changes in AI-directed functional connectivity patterns in PTSD patients. Many within- and between-network causal paths were found to be less influenced by the AI after effective therapy. Insular influences were found to be positively correlated with re-experiencing symptoms, while they were linked with a stronger default mode network (DMN) and more unstable central executive network (CEN) connectivity.

Conclusion

We showed that directed changes in AI signaling to the DMN and CEN at rest may underlie the degree of re-experiencing symptoms in PTSD. A positive response to treatment further induced changes in network-to-network anticorrelated patterns. Such findings may guide targeted neuromodulation strategies in PTSD patients not suitably improved by conventional treatment.

Decoding Activity in Broca's Area Predicts the Occurrence of Auditory Hallucinations Across Subjects

BACKGROUND

Functional magnetic resonance imaging (fMRI) capture aims at detecting auditory-verbal hallucinations (AVHs) from continuously recorded brain activity. Establishing efficient capture methods with low computational cost that easily generalize between patients remains a key objective in precision psychiatry. To address this issue, we developed a novel automatized fMRI-capture procedure for AVHs in patients with schizophrenia (SCZ).

METHODS

We used a previously validated but labor-intensive personalized fMRI-capture method to train a linear classifier using machine learning techniques. We benchmarked the performances of this classifier on 2320 AVH periods versus resting-state periods obtained from SCZ patients with frequent symptoms (n = 23). We characterized patterns of blood oxygen level-dependent activity that were predictive of AVH both within and between subjects. Generalizability was assessed with a second independent sample gathering 2000 AVH labels (n = 34 patients with SCZ), while specificity was tested with a nonclinical control sample performing an auditory imagery task (840 labels, n = 20).

RESULTS

Our between-subject classifier achieved high decoding accuracy (area under the curve = 0.85) and discriminated AVH from rest and verbal imagery. Optimizing the parameters on the first schizophrenia dataset and testing its performance on the second dataset led to an out-of-sample area under the curve of 0.85 (0.88 for the converse test). We showed that AVH detection critically depends on local blood oxygen level-dependent activity patterns within Broca's area.

CONCLUSIONS

Our results demonstrate that it is possible to reliably detect AVH states from fMRI blood oxygen level-dependent signals in patients with SCZ using a multivariate decoder without performing complex preprocessing steps. These findings constitute a crucial step toward brain-based treatments for severe drug-resistant hallucinations.

"Like I'm Talking to a Real Person": Exploring the Meaning of Transference for the Use and Design of AI-Based Applications in Psychotherapy

AI-enabled virtual and robot therapy is increasingly being integrated into psychotherapeutic practice, supporting a host of emotional, cognitive, and social processes in the therapeutic encounter. Given the speed of research and development trajectories of AI-enabled applications in psychotherapy and the practice of mental healthcare, it is likely that therapeutic chatbots, avatars, and socially assistive devices will soon translate into clinical applications much more broadly. While AI applications offer many potential opportunities for psychotherapy, they also raise important ethical, social, and clinical questions that have not yet been adequately considered for clinical practice. In this article, we begin to address one of these considerations: the role of transference in the psychotherapeutic relationship. Drawing on Karen Barad’s conceptual approach to theorizing human–non-human relations, we show that the concept of transference is necessarily reconfigured within AI-human psychotherapeutic encounters. This has implications for understanding how AI-driven technologies introduce changes in the field of traditional psychotherapy and other forms of mental healthcare and how this may change clinical psychotherapeutic practice and AI development alike. As more AI-enabled apps and platforms for psychotherapy are developed, it becomes necessary to re-think AI-human interaction as more nuanced and richer than a simple exchange of information between human and nonhuman actors alone.

Detection of differentiated changes in gray matter in children with progressive hydrocephalus and chronic compensated hydrocephalus using voxel-based morphometry and machine learning

Currently, no neuroimaging study has reported the detection of specific imaging biomarkers that distinguish the progressive hydrocephalus (PH) and chronic compensated hydrocephalus (CH). Our main focus is to evaluate the different structural changes in classifying the two types of hydrocephalus children. Twenty-two children with hydrocephalus (12 PHs and 10 CHs) and 30 age-matched healthy controls were enrolled and the T1-weighted imaging was collected in the study. A customized voxel-based morphometry (VBM) approach and support vector machine (SVM) were combined to investigate the structural changes and group classification. Comparing with the controls and CH, PH groups invariably showed a significant decrease of GM volume in the bilateral hippocampus/parahippocampus, insula, and motor-related areas. SVM applied to the GM volumes of bilateral hippocampus/parahippocampus, insula, and motor-related areas correctly identified hydrocephalus children from normal controls with a statistically significant accuracy of 88.46% (p ≤ .001). In addition, SVM applied to GM volumes of the same regions correctly identified PH from CH with a statistically significant accuracy of 77.27% (p ≤ .009). Using VBM analysis, we characterized and visualized the GM changes in children with hydrocephalus. Machine learning results further confirmed that a significant decrease of the bilateral hippocampus/parahippocampus, insula, and motor-related GM volume can serve as a specific neuroimaging index to distinguish the children with PH from the children with CH and controls at individual. The findings could help to aid the identification of individuals with PH in clinical practice.

Identifying a neuroanatomical signature of schizophrenia, reproducible across sites and stages, using machine learning with structured sparsity

OBJECTIVE

Structural MRI (sMRI) increasingly offers insight into abnormalities inherent to schizophrenia. Previous machine learning applications suggest that individual classification is feasible and reliable and, however, is focused on the predictive performance of the clinical status in cross-sectional designs, which has limited biological perspectives. Moreover, most studies depend on relatively small cohorts or single recruiting site. Finally, no study controlled for disease stage or medication's effect. These elements cast doubt on previous findings' reproducibility.

METHOD

We propose a machine learning algorithm that provides an interpretable brain signature. Using large datasets collected from 4 sites (276 schizophrenia patients, 330 controls), we assessed cross-site prediction reproducibility and associated predictive signature. For the first time, we evaluated the predictive signature regarding medication and illness duration using an independent dataset of first-episode patients.

RESULTS

Machine learning classifiers based on neuroanatomical features yield significant intersite prediction accuracies (72%) together with an excellent predictive signature stability. This signature provides a neural score significantly correlated with symptom severity and the extent of cognitive impairments. Moreover, this signature demonstrates its efficiency on first-episode psychosis patients (73% accuracy).

CONCLUSION

These results highlight the existence of a common neuroanatomical signature for schizophrenia, shared by a majority of patients even from an early stage of the disorder.

Prediction of activation patterns preceding hallucinations in patients with schizophrenia using machine learning with structured sparsity

Despite significant progress in the field, the detection of fMRI signal changes during hallucinatory events remains difficult and time-consuming. This article first proposes a machine-learning algorithm to automatically identify resting-state fMRI periods that precede hallucinations versus periods that do not. When applied to whole-brain fMRI data, state-of-the-art classification methods, such as support vector machines (SVM), yield dense solutions that are difficult to interpret. We proposed to extend the existing sparse classification methods by taking the spatial structure of brain images into account with structured sparsity using the total variation penalty. Based on this approach, we obtained reliable classifying performances associated with interpretable predictive patterns, composed of two clearly identifiable clusters in speech-related brain regions. The variation in transition-to-hallucination functional patterns not only from one patient to another but also from one occurrence to the next (e.g., also depending on the sensory modalities involved) appeared to be the major difficulty when developing effective classifiers. Consequently, second, this article aimed to characterize the variability within the prehallucination patterns using an extension of principal component analysis with spatial constraints. The principal components (PCs) and the associated basis patterns shed light on the intrinsic structures of the variability present in the dataset. Such results are promising in the scope of innovative fMRI-guided therapy for drug-resistant hallucinations, such as fMRI-based neurofeedback.