L. Elliot Hong

(1) Background: The correlations between brain connectivity abnormality and psychiatric disorders have been continuously investigated and progressively recognized. Brain connectivity signatures are becoming exceedingly useful for identifying patients, monitoring mental health disorders, and treatment. By using electroencephalography (EEG)-based cortical source localization along with energy landscape analysis techniques, we can statistically analyze transcranial magnetic stimulation (TMS)-invoked EEG signals, for obtaining connectivity among different brain regions at a high spatiotemporal resolution. (2) Methods: In this study, we analyze EEG-based source localized alpha wave activity in response to TMS administered to three locations, namely, the left motor cortex (49 subjects), left prefrontal cortex (27 subjects), and the posterior cerebellum, or vermis (27 subjects) by using energy landscape analysis techniques to uncover connectivity signatures. We then perform two sample t-tests and use the (5 × 10⁻⁵) Bonferroni corrected p-valued cases for reporting six reliably stable signatures. (3) Results: Vermis stimulation invoked the highest number of connectivity signatures and the left motor cortex stimulation invoked a sensorimotor network state. In total, six out of 29 reliable, stable connectivity signatures are found and discussed. (4) Conclusions: We extend previous findings to localized cortical connectivity signatures for medical applications that serve as a baseline for future dense electrode studies.

Humans are remarkably skilled at listening to one speaker out of an acoustic mixture of several speech sources. Two speakers are easily segregated, even without binaural cues, but the neural mechanisms underlying this ability are not well understood. One possibility is that early cortical processing performs a spectro-temporal decomposition of the acoustic mixture, allowing the attended speech to be reconstructed via optimally weighted recombinations that discount spectro-temporal regions where sources heavily overlap. Using human magnetoencephalography (MEG) responses to a two-talker mixture, we show evidence for an alternative possibility, in which early, active segregation occurs even for strongly spectro-temporally overlapping regions. Early (∼70 ms) responses to non-overlapping spectro-temporal features are seen for both talkers. When competing talkers’ spectro-temporal features mask each other, the individual representations persist, but they occur with a ∼20 ms delay. This suggests that the auditory cortex recovers acoustic features that are masked in the mixture, even if they occurred in the ignored speech. The existence of such noise-robust cortical representations, of features present in attended as well as ignored speech, suggests an active cortical stream segregation process, which could explain a range of behavioral effects of ignored background speech.

Summary: During speech perception, a central task of the auditory cortex is to analyze complex acoustic patterns to allow detection of the words that encode a linguistic message. It is generally thought that this process includes at least one intermediate, phonetic, level of representations [1–6], localized bilaterally in the superior temporal lobe [7–10]. Phonetic representations reflect a transition from acoustic to linguistic information, classifying acoustic patterns into linguistically meaningful units, which can serve as input to mechanisms that access abstract word representations [11–13]. While recent research has identified neural signals arising from successful recognition of individual words in continuous speech [14–17], no explicit neurophysiological signal has been found demonstrating the transition from acoustic/phonetic to symbolic, lexical representations. Here we report a response reflecting the incremental integration of phonetic information for word identification, dominantly localized to the left temporal lobe. The short response latency, approximately 110 ms relative to phoneme onset, suggests that phonetic information is used for lexical processing as soon as it becomes available. Responses also tracked word boundaries, confirming previous reports of immediate lexical segmentation [18,19]. These new results were further investigated using a cocktail-party paradigm [20,21] in which participants listened to a mix of two talkers, attending to one and ignoring the other. Analysis indicates neural lexical processing of only the attended, but not the unattended, speech stream. Thus, while responses to acoustic features reflect attention through selective amplification of attended speech, responses consistent with a lexical processing model reveal categorically selective processing.

Big data initiatives such as the Enhancing NeuroImaging Genetics through Meta-Analysis consortium (ENIGMA), combine data collected by independent studies worldwide to achieve more generalizable estimates of effect sizes and more reliable and reproducible outcomes. Such efforts require harmonized image analyses protocols to extract phenotypes consistently. This harmonization is particularly challenging for resting state fMRI due to the wide variability of acquisition protocols and scanner platforms; this leads to site-to-site variance in quality, resolution and temporal signal-to-noise ratio (tSNR). An effective harmonization should provide optimal measures for data of different qualities. We developed a multi-site rsfMRI analysis pipeline to allow research groups around the world to process rsfMRI scans in a harmonized way, to extract consistent and quantitative measurements of connectivity and to perform coordinated statistical tests. We used the single-modality ENIGMA rsfMRI preprocessing pipeline based on modelfree Marchenko-Pastur PCA based denoising to verify and replicate resting state network heritability estimates. We analyzed two independent cohorts, GOBS (Genetics of Brain Structure) and HCP (the Human Connectome Project), which collected data using conventional and connectomics oriented fMRI protocols, respectively. We used seed-based connectivity and dual-regression approaches to show that the rsfMRI signal is consistently heritable across twenty major functional network measures. Heritability values of 20-40% were observed across both cohorts.

Big data initiatives such as the Enhancing NeuroImaging Genetics through Meta-Analysis consortium (ENIGMA), combine data collected by independent studies worldwide to achieve more accurate estimates of effect sizes and more reliable and reproducible outcomes. Such efforts require harmonized analyses protocols to consistently extract phenotypes. Even so, challenges include wide variability of fMRI protocols and scanner platforms; this leads to site-to-site variance in quality, resolution and temporal signal-to-noise ratio (tSNR). An effective harmonization should provide optimal measures for data of different qualities. We developed a multi-site rsfMRI analysis pipeline to allow research groups around the world to process rsfMRI scans in a harmonized way, to extract consistent and quantitative measurements of connectivity and to perform coordinated statistical tests. We used the single-modality ENIGMA rsfMRI pipeline based on model-free Marchenko-Pastor PCA based denoising to verify and replicate findings of significant heritability of measures from resting state networks. We analyzed two independent cohorts, GOBS (Genetics of Brain Structure) and HCP (the Human Connectome Project), which collected data using conventional and connectomics oriented fMRI protocols. We used seed-based connectivity and dual-regression approaches to show that rsfMRI signal is consistently heritable across twenty major functional network measures. Heritability values of 20-40% were observed across both cohorts.

Delta band (1–4 Hz) neuronal responses support the precision and stability of auditory processing, and a deficit in delta band synchrony may be relevant to auditory domain symptoms in schizophrenia patients. Delta band synchronization elicited by a 2.5 Hz auditory steady state response (ASSR) paradigm, along with those from theta (5 Hz), alpha (10 Hz), beta (20 Hz), gamma (40 Hz), and high gamma (80 Hz) frequency ASSR, were compared in 128 patients with schizophrenia, 108 healthy controls, and 55 first-degree relatives (FDR) of patients. Delta band synchronization was significantly impaired in patients compared with controls (F = 18.3, P < .001). There was a significant 2.5 Hz by 40 Hz ASSR interaction (P = .023), arising from a greater reduction of 2.5 Hz ASSR than of 40 Hz ASSR, in patients compared with controls. Greater deficit in delta ASSR was associated with auditory perceptual abnormality (P = .007) and reduced verbal working memory (P < .001). Gamma frequency ASSR impairment was also significant but more modest (F = 8.7, P = .004), and this deficit was also present in FDR (P = .022). The ability to sustain delta band oscillation entrainment in the auditory pathway is significantly reduced in schizophrenia patients and appears to be clinically relevant.

We investigated in vivo neurochemical markers reflective of neuronal health and glial activation to determine if these could yield clues regarding the reduced fractional anisotropy (FA) of white matter and accelerated decline of FA with age in schizophrenia. Participants with schizophrenia and healthy controls completed diffusion tensor imaging to assess FA and proton magnetic resonance spectroscopy to assess neurochemical metabolites in the same frontal region. Frontal FA was significantly lower in the schizophrenia and declined more rapidly with age compared with the healthy control group. In both groups, N-acetylaspartate (NAA), a putative marker of neuronal integrity, and glutamate declined with age, and this decline was stronger in patients. Myo-inositol, a marker of glial cells, was negatively related to FA in both groups. The relationship between FA and age remained significant in schizophrenia even when controlling for all metabolites. The relationships of FA, NAA and myo-inositol to age appear to be independent of one another. The relationship between FA and myo-inositol was independently present in both patients and controls, even after controlling for age, indicating a potential general effect of neuroinflammation on white matter microstructure. Further studies are warranted to determine the underlying mechanism driving the accelerated FA decline with age in schizophrenia.

Depression is highly prevalent in patients with schizophrenia and is associated with significant clinical consequences, but there is no known biomarker for depression in schizophrenia. One of the putative neurochemical biomarkers for depression in major depressive disorder is reduced cerebral concentration of myo-Inositol. We examined whether myo-Inositol levels provide a potential marker for depressive symptoms in schizophrenia similar to that in major depressive disorder and are informative regarding causal biological pathways underlying both depression and schizophrenia. We used proton magnetic resonance spectroscopy to examine myo-Inositol levels in the anterior cingulate cortex (ACC) in 59 schizophrenia spectrum disorder patients and 69 matched community comparison participants. Participants completed the Maryland Trait and State Depression (MTSD) scale to measure symptoms of depression experienced around time of assessment (’State’ subscale) and longitudinally (‘Trait’ subscale). Myo-Inositol in the ACC was negatively correlated with MTSD-trait scores in both patients (rho=−0.336, p=0.009) and community comparison samples (rho=−0.328, p=0.006). Furthermore, patients with a diagnosis of schizoaffective disorder or a history of at least one major depressive episode had lower levels of myo-Inositol compared to schizophrenia patients without a current or past affective diagnosis (p=0.012). Since reduced brain myo-Inositol is associated with major depressive disorder, myo-Inositol may be a biochemical marker of depressive mood symptoms across diagnostic boundaries. If confirmed, this finding may aid investigation of the pathophysiology and therapeutics of depression common between depression, schizophrenia, and other psychiatric diagnoses.

Background: Brain-derived neurotrophic factor (BDNF) protein levels decline in the brain during senescence and are also shown to be reduced in schizophrenia patients. BDNF is present in both the gray and white matters of the brain. It is unclear whether BDNF abnormalities in schizophrenia are specific to gray and/or white matter. Objective: We hypothesized that the age-related BDNF decline is abnormal and contributes to the reduced BDNF in schizophrenia. Methods: We tested this hypothesis by measuring BDNF protein levels in postmortem gray and white matter, using the prefrontal cortex (PFC) and the genu of the corpus callosum as regions of interests, from 20 schizophrenia patients and 20 matched nonpsychiatric controls. Samples were selected across the adult lifespan - from 20 to 80 years of age. Results: PFC gray matter BDNF protein levels were significantly lower in older age in both nonpsychiatric comparisons and patients, while BDNF in white matter did not decrease significantly with age in either group. PFC BDNF was linearly lower from 20 to 80 years of age in nonpsychiatric comparisons. In schizophrenia, the age effect was similarly linear in younger patients but a decline did not occur in older patients. Conclusion: PFC BDNF does not follow a normative linear age effect in schizophrenia patients as they grow older, which may represent a ‘floor effect' due to earlier decline or a survivor cohort of older patient donors who are less susceptible to a schizophrenia-related pathological aging process.

Integration of diverse synaptic inputs is a basic neuronal operation that relies on many neurocomputational principles, one of which is neural summation. However, we lack empirical understanding of neuronal summation in the human brains in vivo. Here, we explored the effect of neural summation on the motor cortex using two subthreshold pulses of transcranial magnetic stimulation (TMS), each with intensities ranging from 60 to 95 % of the resting motor threshold (RMT) and interstimulus interval (ISI) varying from 1 to 25 ms. We found that two subthreshold TMS pulses can produce suprathreshold motor response when ISIs were less than 10 ms, most prominent at 1, 1.5 and 3 ms. This facilitatory, above-threshold response was evident when the intensity of the subthreshold pulses was above 80 % of RMT but was absent as the intensity was 70 % or below. Modeling of the summation data across intensity suggested that they followed an exponential function with excellent model fitting. Understanding the constraints for inducing summation of subthreshold stimulations to generate above-threshold response may have implications in modeling neural operations and potential clinical applications.