Extending Subcortical EEG Responses to Continuous Speech to the Sound-Field.

The auditory brainstem response (ABR) is a valuable clinical tool for objective hearing assessment, which is conventionally detected by averaging neural responses to thousands of short stimuli. Progressing beyond these unnatural stimuli, brainstem responses to continuous speech presented via earphones have been recently detected using linear temporal response functions (TRFs). Here, we extend earlier studies by measuring subcortical responses to continuous speech presented in the sound-field, and assess the amount of data needed to estimate brainstem TRFs. Electroencephalography (EEG) was recorded from 24 normal hearing participants while they listened to clicks and stories presented via earphones and loudspeakers. Subcortical TRFs were computed after accounting for non-linear processing in the auditory periphery by either stimulus rectification or an auditory nerve model. Our results demonstrated that subcortical responses to continuous speech could be reliably measured in the sound-field. TRFs estimated using auditory nerve models outperformed simple rectification, and 16 minutes of data was sufficient for the TRFs of all participants to show clear wave V peaks for both earphones and sound-field stimuli. Subcortical TRFs to continuous speech were highly consistent in both earphone and sound-field conditions, and with click ABRs. However, sound-field TRFs required slightly more data (16 minutes) to achieve clear wave V peaks compared to earphone TRFs (12 minutes), possibly due to effects of room acoustics. By investigating subcortical responses to sound-field speech stimuli, this study lays the groundwork for bringing objective hearing assessment closer to real-life conditions, which may lead to improved hearing evaluations and smart hearing technologies.

Predictors for estimating subcortical EEG responses to continuous speech.

Perception of sounds and speech involves structures in the auditory brainstem that rapidly process ongoing auditory stimuli. The role of these structures in speech processing can be investigated by measuring their electrical activity using scalp-mounted electrodes. However, typical analysis methods involve averaging neural responses to many short repetitive stimuli that bear little relevance to daily listening environments. Recently, subcortical responses to more ecologically relevant continuous speech were detected using linear encoding models. These methods estimate the temporal response function (TRF), which is a regression model that minimises the error between the measured neural signal and a predictor derived from the stimulus. Using predictors that model the highly non-linear peripheral auditory system may improve linear TRF estimation accuracy and peak detection. Here, we compare predictors from both simple and complex peripheral auditory models for estimating brainstem TRFs on electroencephalography (EEG) data from 24 participants listening to continuous speech. We also investigate the data length required for estimating subcortical TRFs, and find that around 12 minutes of data is sufficient for clear wave V peaks (>3 dB SNR) to be seen in nearly all participants. Interestingly, predictors derived from simple filterbank-based models of the peripheral auditory system yield TRF wave V peak SNRs that are not significantly different from those estimated using a complex model of the auditory nerve, provided that the nonlinear effects of adaptation in the auditory system are appropriately modelled. Crucially, computing predictors from these simpler models is more than 50 times faster compared to the complex model. This work paves the way for efficient modelling and detection of subcortical processing of continuous speech, which may lead to improved diagnosis metrics for hearing impairment and assistive hearing technology.

Predicting EEG Responses to Attended Speech via Deep Neural Networks for Speech.

Attending to the speech stream of interest in multi-talker environments can be a challenging task, particularly for listeners with hearing impairment. Research suggests that neural responses assessed with electroencephalography (EEG) are modulated by listener's auditory attention, revealing selective neural tracking (NT) of the attended speech. NT methods mostly rely on hand-engineered acoustic and linguistic speech features to predict the neural response. Only recently, deep neural network (DNN) models without specific linguistic information have been used to extract speech features for NT, demonstrating that speech features in hierarchical DNN layers can predict neural responses throughout the auditory pathway. In this study, we go one step further to investigate the suitability of similar DNN models for speech to predict neural responses to competing speech observed in EEG. We recorded EEG data using a 64-channel acquisition system from 17 listeners with normal hearing instructed to attend to one of two competing talkers. Our data revealed that EEG responses are significantly better predicted by DNN-extracted speech features than by hand-engineered acoustic features. Furthermore, analysis of hierarchical DNN layers showed that early layers yielded the highest predictions. Moreover, we found a significant increase in auditory attention classification accuracies with the use of DNN-extracted speech features over the use of hand-engineered acoustic features. These findings open a new avenue for development of new NT measures to evaluate and further advance hearing technology.

Early prediction of mental health problems following military deployment: Integrating pre- and post-deployment factors in neural network models.

Military personnel deployed to war zones are at increased risk of mental health problems such as posttraumatic stress disorder (PTSD) or depression. Early pre- or post-deployment identification of those at highest risk of such problems is crucial to target intervention to those in need. However, sufficiently accurate models predicting objectively assessed mental health outcomes have not been put forward. In a sample consisting of all Danish military personnel who deployed to war zones for the first (N = 27,594), second (N = 11,083) and third (N = 5,161) time between 1992 and 2013, we apply neural networks to predict psychiatric diagnoses or use of psychotropic medicine in the years following deployment. Models are based on pre-deployment registry data alone or on pre-deployment registry data in combination with post-deployment questionnaire data on deployment experiences or early post-deployment reactions. Further, we identified the most central predictors of importance for the first, second, and third deployment. Models based on pre-deployment registry data alone had lower accuracy (AUCs ranging from 0.61 (third deployment) to 0.67 (first deployment)) than models including pre- and post-deployment data (AUCs ranging from 0.70 (third deployment) to 0.74 (first deployment)). Age at deployment, deployment year and previous physical trauma were important across deployments. Post-deployment predictors varied across deployments but included deployment exposures as well as early post-deployment symptoms. The results suggest that neural network models combining pre- and early post-deployment data can be utilized for screening tools that identify individuals at risk of severe mental health problems in the years following military deployment.

Resting-state EEG functional connectivity predicts post-traumatic stress disorder subtypes in veterans.

Objective. Post-traumatic stress disorder (PTSD) is highly heterogeneous, and identification of quantifiable biomarkers that could pave the way for targeted treatment remains a challenge. Most previous electroencephalography (EEG) studies on PTSD have been limited to specific handpicked features, and their findings have been highly variable and inconsistent. Therefore, to disentangle the role of promising EEG biomarkers, we developed a machine learning framework to investigate a wide range of commonly used EEG biomarkers in order to identify which features or combinations of features are capable of characterizing PTSD and potential subtypes.Approach. We recorded 5 min of eyes-closed and 5 min of eyes-open resting-state EEG from 202 combat-exposed veterans (53% with probable PTSD and 47% combat-exposed controls). Multiple spectral, temporal, and connectivity features were computed and logistic regression, random forest, and support vector machines with feature selection methods were employed to classify PTSD. To obtain robust results, we performed repeated two-layer cross-validation to test on an entirely unseen test set.Main results. Our classifiers obtained a balanced test accuracy of up to 62.9% for predicting PTSD patients. In addition, we identified two subtypes within PTSD: one where EEG patterns were similar to those of the combat-exposed controls, and another that were characterized by increased global functional connectivity. Our classifier obtained a balanced test accuracy of 79.4% when classifying this PTSD subtype from controls, a clear improvement compared to predicting the whole PTSD group. Interestingly, alpha connectivity in the dorsal and ventral attention network was particularly important for the prediction, and these connections were positively correlated with arousal symptom scores, a central symptom cluster of PTSD.Significance. Taken together, the novel framework presented here demonstrates how unsupervised subtyping can delineate heterogeneity and improve machine learning prediction of PTSD, and may pave the way for better identification of quantifiable biomarkers.

Forecasting military mental health in a complete sample of Danish military personnel deployed between 1992-2013.

OBJECTIVE: Mental health problems (MHP) are a relatively common consequence of deployment to war zones. Early identification of those at risk of post-deployment MHP would improve prevention efforts. However, screening instruments based on linear models have not been successful. Machine learning (ML) has shown promise for providing the methodological frame for better prognostic models. METHODS: The study population was all Danish military personnel deployed for the first time between January 1, 1992 and December 31, 2013. From extensive registry data, 21 pre- or at-deployment predictors comprising early adversity, social, clinical and demographic variables were used to predict psychiatric contacts (psychiatric diagnosis and/or use of psychotropic medicine) occurring within 6.5 years after homecoming. Four supervised ML methods (penalized logistic regression, random forests, support vector machines and gradient boosting machines) were compared in ability to classify those with high risk of post-deployment MHP and those without. RESULTS: Of 27594 subjects, 2175 (8%) had a psychiatric contact. All four ML methods applied had performances well above chance (Area under the Receiver-operating Curve 0.62-0.68). Positive predictive value for the best model was 0.16. A range of pre-deployment factors were found to be predictive of post-deployment psychiatric contacts. CONCLUSIONS: ML methods can be useful in early identification of soldiers with high risk of MPH in the years following their first deployment. However, performances were modest and positive predictive values were low, limiting the applicability of the models for pre-deployment screening. Future studies should include neurobiological data and deployment experiences to increase accuracy of the models.

Applicability of an Automated Model and Parameter Selection in the Prediction of Screening-Level PTSD in Danish Soldiers Following Deployment: Development Study of Transferable Predictive Models Using Automated Machine Learning.

BACKGROUND: Posttraumatic stress disorder (PTSD) is a relatively common consequence of deployment to war zones. Early postdeployment screening with the aim of identifying those at risk for PTSD in the years following deployment will help deliver interventions to those in need but have so far proved unsuccessful. OBJECTIVE: This study aimed to test the applicability of automated model selection and the ability of automated machine learning prediction models to transfer across cohorts and predict screening-level PTSD 2.5 years and 6.5 years after deployment. METHODS: Automated machine learning was applied to data routinely collected 6-8 months after return from deployment from 3 different cohorts of Danish soldiers deployed to Afghanistan in 2009 (cohort 1, N=287 or N=261 depending on the timing of the outcome assessment), 2010 (cohort 2, N=352), and 2013 (cohort 3, N=232). RESULTS: Models transferred well between cohorts. For screening-level PTSD 2.5 and 6.5 years after deployment, random forest models provided the highest accuracy as measured by area under the receiver operating characteristic curve (AUC): 2.5 years, AUC=0.77, 95% CI 0.71-0.83; 6.5 years, AUC=0.78, 95% CI 0.73-0.83. Linear models performed equally well. Military rank, hyperarousal symptoms, and total level of PTSD symptoms were highly predictive. CONCLUSIONS: Automated machine learning provided validated models that can be readily implemented in future deployment cohorts in the Danish Defense with the aim of targeting postdeployment support interventions to those at highest risk for developing PTSD, provided the cohorts are deployed on similar missions.

Anhedonia and emotional numbing in treatment-seeking veterans: behavioural and electrophysiological responses to reward.

Background: Anhedonia is a common symptom following exposure to traumatic stress and a feature of the PTSD diagnosis. In depression research, anhedonia has been linked to deficits in reward functioning, reflected in behavioural and neural responses. Such deficits following exposure to trauma, however, are not well understood. Objective: The current study aims to estimate the associations between anhedonia, PTSD symptom-clusters and behavioural and electrophysiological responses to reward. Methods: Participants (N = 61) were recruited among Danish treatment-seeking veterans at the Department of Military Psychology in the Danish Defence. Before entering treatment, participants were screened with symptom measurement instruments and participated in a joint behavioural-electrophysiological experiment. The experimental paradigm consisted of a signal-detection task aimed at assessing reward-driven learning. Simultaneous electrophysiological-recordings were analysed to evaluate neural responses upon receiving reward, as indicated by the Feedback-Related Negativity (FRN) component. Result: Anhedonia as conceptualized in depression correlated with behavioural learning (r = -0.28, p = .032). Neither anhedonia nor behavioural learning correlated with FRN. However, the anhedonia symptom cluster of PTSD did correlate with FRN (r = 0.29, p = .023). Extending upon this in an exploratory analysis, the specific PTSD-symptom emotional numbing was found to correlate moderately with FRN (r = 0.38, p = .003). Conclusion: The present data suggest that anhedonia in trauma-exposed individuals is related to the anticipatory aspect of reward, whereas the neural consummatory reward response seems unlinked. Interestingly, emotional numbing in the same population is related to the consummatory phase of reward, correlating with the FRN response. This suggests that anhedonia and emotional numbing in response to trauma might pertain to different phases of reward processing.

Planteamiento: La anhedonia es un síntoma frecuente después de la exposición al estrés traumático y una característica del diagnóstico de TEPT. En la investigación de la depresión, la anhedonia se ha relacionado con los déficits en el funcionamiento de la recompensa, que se refleja en las respuestas conductuales y neuronales. Dichos déficits después de la exposición al trauma, sin embargo, no se entienden bien.Objetivo: El presente estudio tiene como objetivo estimar las asociaciones entre la anhedonia, los grupos de síntomas de TEPT y las respuestas conductuales y electrofisiológicas a la recompensa. Métodos: Los participantes (N = 61) fueron reclutados entre los veteranos daneses que buscaban tratamiento en el Departamento de Psicología Militar de la Defensa Danesa. Antes de empezar al tratamiento, los participantes fueron evaluados con instrumentos de medición de síntomas y participaron en un experimento conjunto electrofisiológico-conductual. El paradigma experimental consistió en una tarea de detección de señales destinada a evaluar el aprendizaje basado en recompensas. Simultáneamente se analizaron las grabaciones de EEG para evaluar las respuestas neurales al recibir la recompensa, según lo indicado por el componente de negatividad relacionada con la retroalimentación (FRN, por sus siglas en inglés).Resultados: La anhedonia, tal como se conceptualizó en la depresión, se correlacionó con el aprendizaje conductual (r = −0.28, p = .032). Ni la anhedonia ni el aprendizaje conductual se correlacionaron con la FRN. Sin embargo, el grupo de síntomas de anhedonia del TEPT se correlacionó con el FRN (r = 0.29, p = .023). Sobre la base de esto en un análisis exploratorio, el entumecimiento emocional específico de los síntomas del TEPT se correlacionó moderadamente con la FRN (r = 0.38, p = .003).Conclusiones: Los datos actuales sugieren que la anhedonia en individuos expuestos al trauma está relacionada con el aspecto anticipatorio de la recompensa, mientras que la respuesta de recompensa neural consumada parece desvinculada. Curiosamente, el adormecimiento emocional en la misma población está relacionado con la fase de recompensa consumada, que se correlaciona con la respuesta FRN. Esto sugiere que la anhedonia y el entumecimiento emocional en respuesta al trauma podrían pertenecer a diferentes fases del procesamiento de la recompensa.

Face configuration affects speech perception: Evidence from a McGurk mismatch negativity study.

We perceive identity, expression and speech from faces. While perception of identity and expression depends crucially on the configuration of facial features it is less clear whether this holds for visual speech perception. Facial configuration is poorly perceived for upside-down faces as demonstrated by the Thatcher illusion in which the orientation of the eyes and mouth with respect to the face is inverted (Thatcherization). This gives the face a grotesque appearance but this is only seen when the face is upright. Thatcherization can likewise disrupt visual speech perception but only when the face is upright indicating that facial configuration can be important for visual speech perception. This effect can propagate to auditory speech perception through audiovisual integration so that Thatcherization disrupts the McGurk illusion in which visual speech perception alters perception of an incongruent acoustic phoneme. This is known as the McThatcher effect. Here we show that the McThatcher effect is reflected in the McGurk mismatch negativity (MMN). The MMN is an event-related potential elicited by a change in auditory perception. The McGurk-MMN can be elicited by a change in auditory perception due to the McGurk illusion without any change in the acoustic stimulus. We found that Thatcherization disrupted a strong McGurk illusion and a correspondingly strong McGurk-MMN only for upright faces. This confirms that facial configuration can be important for audiovisual speech perception. For inverted faces we found a weaker McGurk illusion but, surprisingly, no MMN. We also found no correlation between the strength of the McGurk illusion and the amplitude of the McGurk-MMN. We suggest that this may be due to a threshold effect so that a strong McGurk illusion is required to elicit the McGurk-MMN.

Multistage audiovisual integration of speech: dissociating identification and detection.

Speech perception integrates auditory and visual information. This is evidenced by the McGurk illusion where seeing the talking face influences the auditory phonetic percept and by the audiovisual detection advantage where seeing the talking face influences the detectability of the acoustic speech signal. Here, we show that identification of phonetic content and detection can be dissociated as speech-specific and non-specific audiovisual integration effects. To this end, we employed synthetically modified stimuli, sine wave speech (SWS), which is an impoverished speech signal that only observers informed of its speech-like nature recognize as speech. While the McGurk illusion only occurred for informed observers, the audiovisual detection advantage occurred for naïve observers as well. This finding supports a multistage account of audiovisual integration of speech in which the many attributes of the audiovisual speech signal are integrated by separate integration processes.