Search

Reward processing is essential for our mental-health and well-being. Here, we present the development and validation of a scalable fMRI-informed EEG model related to reward processing in the ventral-striatum (VS); a central reward circuit node. Simultaneous EEG/fMRI data were acquired from 17 healthy individuals listening to pleasurable music, and used to construct a one-class regression model for predicting the reward-related VS-BOLD signal using spectro-temporal features from the EEG. Validation analyses, applied on EEG/fMRI data from a different group (N=14), revealed that the EEG model predicted VS-BOLD activation from the simultaneous EEG to a greater extent than a model derived from another anatomical region. The VS-EEG-model was also modulated by musical pleasure and predictive of the VS-BOLD during a monetary reward task, further indicating it functional relevance. These findings provide compelling evidence for the use of a scalable yet precise EEG-only probe of VS-originated reward processing, which could serve for process specific neruo-monitoring and -modulation.

Amygdala dysregulation is core to multiple psychiatric disorders. Real-time fMRI enables Amygdala self-modulation through NeuroFeedback (NF).Despite a surge in Amygdala-NF studies, a systematic quantification of self-modulation is lacking. Amygdala-NF dissemination is further restricted by absence of unifying framework dictating design choices and insufficient understanding of neural changes underlying successful self-modulation.The current meta-analysis of Amygdala-NF literature found that real-time feedback facilitates learned self-modulation more than placebo. Intriguingly, while we found that variability in design choices could be explained by the targeted domain, this was rarely highlighted by authors. Lastly, reanalysis of six fMRI data-sets (n=151), revealed that successful Amygdala down-modulation is coupled with deactivation of posterior insula and Default-Mode-Network major nodes, pointing to regulation related processes.While findings point to Amygdala self-modulation as a learned skill that could modify brain functionality, further placebo-controlled trials are necessary to prove clinical efficacy. We further suggest that studies should explicitly target neuro-behavioral domain, design studies accordingly and include ‘target engagement’ measures. We exemplify this idea through a ‘process-based’ NF approach for PTSD.

Negative and positive valence systems (NVS and PVS) pertain to processing of aversive and rewarding stimuli, respectively. Post-Traumatic Stress Disorder (PTSD) has been typically associated with abnormalities of the NVS, mostly related to heightened threat processing, yet more recent work also suggests deficits in PVS functionality in PTSD, mainly in the form of reduced reward functioning. The current study examined the hypothesis that individuals ability to recover from a potentially traumatic event relies on promoting reward-seeking behaviors (i.e., PVS) alongside diminished threat assessment (i.e., NVS), during the first year following trauma, a critical period for PTSD development or recovery. To do so, we longitudinally tracked behavioral and neural responses among 171 adult civilians with early post-traumatic stress symptoms at 1-, 6- and 14-months following trauma exposure (TP1, TP2, and TP3, respectively). At each time-point, participants played a naturalistic game encompassing dynamic provocation of risk-taking, punishments and rewards in an fMRI setting. Results showed that greater amygdala activation and functional connectivity with the lateral orbitofrontal cortex (lOFC) in response to punishments (i.e., hyperactive NVS) at TP1 were associated with more severe post-traumatic stress symptoms at both TP1 and TP3 (but not at TP2), and specifically with more hyperarousal and intrusion symptoms. On the other hand, decreased ventral striatum (VS) activity and functional connectivity with the ventromedial prefrontal cortex (vmPFC) in response to rewards (i.e., hypoactive PVS) at TP1 were associated with more severe post-traumatic stress symptoms at TP3 (but not at TP1 or TP2), and specifically with more avoidance symptoms. Explainable machine learning revealed the primacy of the PVS over the NVS at TP1 in predicting PTSD symptom development from TP1 to TP3. Behaviorally, fewer risky choices in the task were associated with more severe symptoms at TP1, but not at TP2 or TP3. Finally, an integrative exploratory analysis revealed that reduction in risky choices in the task (from TP1 to TP2) moderated the relation between NVS hyperactivity at TP1 and symptom severity at TP3. Altogether, our results support the idea that trauma exposure might alter both NVS and PVS processing. While NVS presents early heightened saliency processing in the immediate aftermath of trauma, early PVS only affects the long-term outcome of traumatic stress. These insights inform possible mechanism-driven therapeutic strategies for PTSD, addressing not only negative but also positive valence processing.

Real-time fMRI neurofeedback is an increasingly popular neuroimaging technique that allows an individual to gain control over his/her own brain signals, which can lead to improvements in behavior in healthy participants as well as to improvements of clinical symptoms in patient populations. However, a considerably large ratio of participants undergoing neurofeedback training do not learn to control their own brain signals and, consequently, do not benefit from neurofeedback interventions, which limits clinical efficacy of neurofeedback interventions. As neurofeedback success varies between studies and participants, it is important to identify factors that might influence neurofeedback success. Here, for the first time, we employed a big data machine learning approach to investigate the influence of 20 different design-specific (e.g. activity vs. connectivity feedback), region of interest-specific (e.g. cortical vs. subcortical) and subject-specific factors (e.g. age) on neurofeedback performance and improvement in 608 participants from 28 independent experiments.With a classification accuracy of 60% (considerably different from chance level), we identified two factors that significantly influenced neurofeedback performance: Both the inclusion of a pre-training no-feedback run before neurofeedback training and neurofeedback training of patients as compared to healthy participants were associated with better neurofeedback performance. The positive effect of pre-training no-feedback runs on neurofeedback performance might be due to the familiarization of participants with the neurofeedback setup and the mental imagery task before neurofeedback training runs. Better performance of patients as compared to healthy participants might be driven by higher motivation of patients, higher ranges for the regulation of dysfunctional brain signals, or a more extensive piloting of clinical experimental paradigms. Due to the large heterogeneity of our dataset, these findings likely generalize across neurofeedback studies, thus providing guidance for designing more efficient neurofeedback studies specifically for improving clinical neurofeedback-based interventions. To facilitate the development of data-driven recommendations for specific design details and subpopulations the field would benefit from stronger engagement in Open Science and data sharing.

Genetic counseling (GC) for hereditary breast and ovarian cancer is available mainly in academic settings. Despite equal risk, most low income public hospital patients remain unaware and untested. Remote counseling may be a solution, but research has been limited to phone counseling for insured patients. Our study compares in-person, phone, and video conference GC among high-risk patients in 3 public hospitals to determine the comparative effectiveness of GC delivered across modes with regard to patients’ knowledge, cancer distress, decisional conflict, perceived stress, risk perception, satisfaction, and recall. We also assessed whether patients have a preference for counseling mode and how that affects outcomes. This report describes the study design and lessons learned regarding recruitment. We conducted a multicenter partially randomized preference noninferiority trial with English-, Spanish-, and Cantonese-speaking patients assigned by randomization or patients´ preference to one of the three GC modes. High-risk patients were identified using a family history screener in clinics or by physician referral. Study staff verified risk by phone, invited participation, conducted informed consent, and administered a baseline survey. Enrollees were asked whether they could be randomized or if they preferred one GC mode. They were then given a GC appointment and called again within 2 weeks of counseling for a follow-up survey. Power calculations required 270 randomized patients. A total of 23,401 screener forms yielded 824 likely to be high-risk; 656 completed baseline surveys. Race/ethnic composition was 40% Latinx, 25% white, 19% African American, and 8% Asian. Of these, 531 were counseled, and 505 completed final surveys (283 from randomized patients). The majority (64%) of non-randomized patients chose counseling by phone, 33% chose in person, 3% chose video. • At every step, participation exceeded our projections, showing that diverse low-income patients were interested in participating in research that they deemed relevant. • Our greatest recruitment challenges were due more to settings than to patients. Collection of screeners varied greatly by month and/or clinic. Oncologists valued the risk services offered by the study, but intensive engagement was necessary with front-line staff/supervisors because of their job demands. • Partial randomization functioned well. Prior studies showed that many high-risk women refuse randomization for GC. Adding a preference arm necessitated a larger sample, but greater inclusiveness yields more generalizable findings. • Recruitment of Chinese-speaking patients was low (2.5%) due largely to structural barriers which we continue to explore. Practice-based safety net research presents numerous challenges that require close partnerships, extensive planning, and highly skilled staff capable of sensitive personnel engagement. The work is rewarded by real-world findings, the sine qua non in efforts to eliminate cancer disparities. Citation Format: Claudia Guerra, Robin Lee, Susan L Stewart, Celia Kaplan, Galen Joseph, Janice Tsoh, Niharika Dixit, Heather Cedermaz, Jin Kim, Jane Campbell, Lily X Wang, Amal Khoury, Cindy Hellman-Wylie, Rena J Pasick. Extending the reach of genetic counseling to the safety net: Study design and recruitment challenges of a randomized trial [abstract]. In: Proceedings of the Twelfth AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2019 Sep 20-23; San Francisco, CA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2020;29(6 Suppl_2):Abstract nr A034.

Neurofeedback training has been shown to influence behavior in healthy participants as well as to alleviate clinical symptoms in neurological, psychosomatic, and psychiatric patient populations. However, many real-time fMRI neurofeedback studies report large interindividual differences in learning success. The factors that cause this vast variability between participants remain unknown and their identification could enhance treatment success. Thus, here we employed a meta-analytic approach including data from 24 different neurofeedback studies with a total of 401 participants, including 140 patients, to determine whether levels of activity in target brain regions during pre-training functional localizer or no-feedback runs (i.e., self-regulation in the absence of neurofeedback) could predict neurofeedback learning success. We observed a slightly positive correlation between pre-training activity levels during a functional localizer run and neurofeedback learning success, but we were not able to identify common brain-based success predictors across our diverse cohort of studies. Therefore, advances need to be made in finding robust models and measures of general neurofeedback learning, and in increasing the current study database to allow for investigating further factors that might influence neurofeedback learning.

BackgroundLow hippocampal volume could serve as an early risk factor for Post-traumatic Stress Disorder (PTSD) in interaction with other brain anomalies of developmental origin. One such anomaly may well be a presence of large Cavum Septum Pellucidum (CSP), which has been loosely associated with PTSD. Here, we performed a longitudinal prospective study of recent trauma survivors. We hypothesized that at one-month after trauma exposure, the relation between hippocampal volume and PTSD symptom severity will be moderated by CSP volume, and that this early interaction will account for persistent PTSD symptoms at subsequent time-points.Methods171 adults (87 females, average age=34.22, range=18-65) admitted to a general hospital’s emergency department following a traumatic event, underwent clinical assessment and structural MRI within one-month after trauma. Follow-up clinical evaluations were conducted at six (n=97) and fourteen (n=78) months after trauma. Hippocampus and CSP volumes were measured automatically by FreeSurfer software and verified manually by a neuroradiologist.ResultsAt one-month following trauma, CSP volume significantly moderated the relation between hippocampal volume and PTSD severity (p=0.026), and this interaction further predicted symptom severity at fourteen months post-trauma (p=0.018). Specifically, individuals with smaller hippocampus and larger CSP at one-month post-trauma, showed more severe symptoms at one-and fourteen months following trauma exposure.ConclusionsOur study provides evidence for an early neuroanatomical risk factors for PTSD, which could also predict the progression of the disorder in the year following trauma exposure. Such a simple-to-acquire neuroanatomical signature for PTSD could guide early management, as well as long-term monitoring.Trial RegistrationNeurobehavioral Moderators of Post-traumatic Disease Trajectories. ClinicalTrials.gov registration:NCT03756545.https://clinicaltrials.gov/ct2/show/NCT03756545

Contemporary symptom-based diagnosis of Post-traumatic Stress Disorder (PTSD) largely overlooks related neurobehavioral findings and rely entirely on subjective interpersonal reporting. Previous studies associating objective biomarkers with PTSD have mostly used the disorder’s symptom-based diagnosis as main outcome measure, overlooking the actual clustering and richness of phenotypical features associated with PTSD. Here, we aimed to computationally derive potential neurocognitive biomarkers that could efficiently differentiate PTSD subtypes, based on an observational cohort study of recent trauma survivors. A three-staged semi-unsupervised method (“3C”) was used to categorize trauma survivors based on current PTSD diagnostics, derive clusters of PTSD based on features related to symptom load, and to classify participants’ cluster membership using objective features. A total of 256 features were extracted from psychometrics, cognitive, structural and functional neuroimaging data, obtained from 101 adult civilians (age=34.80±11.95, 51 females) evaluated within a month of trauma exposure. Multi-domain features that best differentiated cluster membership were indicated by using importance analysis, classification trees, and ANOVA. Results revealed that entorhinal and rostral anterior cingulate cortices volumes (structural domain), in-task amygdala’s functional connectivity with the insula and thalamus (functional domain), executive function and cognitive flexibility (cognitive domain) best differentiated between two clusters related to PTSD severity. Cross-validation established the results’ robustness and consistency within this sample. Multi-domain biomarkers revealed by the 3C analytics offer objective classifiers of post-traumatic morbidity shortly following trauma. They also map onto previously documented neurobehavioral PTSD features, supporting the future use of standardized and objective measurements to more precisely identify psychopathology subgroups shortly after trauma.

BackgroundDeep neural networks have revolutionised machine learning, with unparalleled performance in object classification. However, in brain imaging (e.g. fMRI), the direct application of Convolutional Neural Networks (CNN) to decoding subject states or perception from imaging data seems impractical given the scarcity of available data.New methodIn this work we propose a robust method to transfer information from deep learning (DL) features to brain fMRI data with the goal of decoding. By adopting Reduced Rank Regression with Ridge Regularisation we establish a multivariate link between imaging data and the fully connected layer (fc7) of a CNN. We exploit the reconstructed fc7 features by performing an object image classification task on two datasets: one of the largest fMRI databases, taken from different scanners from more than two hundred subjects watching different movie clips, and another with fMRI data taken while watching static images,ResultsThe fc7 features could be significantly reconstructed from the imaging data, and led to significant decoding performance.Comparison with existing methodsThe decoding based on reconstructed fc7 outperformed the decoding based on imaging data alone.ConclusionIn this work we show how to improve fMRI-based decoding benefiting from the mapping between functional data and CNN features. The potential advantage of the proposed method is twofold: the extraction of stimuli representations by means of an automatic procedure (unsupervised) and the embedding of high-dimensional neuroimaging data onto a space designed for visual object discrimination, leading to a more manageable space from dimensionality point of view.