Neural Encoding and Decoding With Distributed Sentence Representations.

Altered functional brain activity in first-episode major depressive disorder treated with electro-acupuncture: A resting-state functional magnetic resonance imaging study

Dual coding of knowledge in the human brain.

A mixed herbal extract as an adjunctive therapy for attention deficit hyperactivity disorder: A randomized placebo-controlled trial

This study aimed to investigate individual differences' effects on brain activity in selected and non-selected avatars for re-engagement. The development of some applications for human-computer interaction has accelerated over the past decade. To develop a human-computer communication system using virtual avatars without losing the user's interest and attention, this study revealed differences in the neural mechanisms underlying the perception of virtual avatars between avatars with which users want to converse again (selected avatars) and those with which they do not (non-selected avatars). Forty-two individuals were recruited; they watched two videos in sequence in which each virtual avatar greeted them, and they then reported which avatars they wanted to talk to again. Meanwhile, brain activities were recorded by functional magnetic resonance imaging (fMRI). After the fMRI recording, the responses to the questionnaires regarding personality traits and avatar impressions were rated. Brain activities were compared along with the score of each personality questionnaire. The results indicated that the left middle temporal gyrus (MTG) was more active in selected compared to non-selected avatars. Furthermore, the brain activities of right superior frontal gyrus (SFG) and the left middle cingulate gyrus (MCG) had a statistically negative correlation with the score of openness in the Ten-Item Personality Inventory (TIPI) in selected avatars. These findings indicate that neural responses during brief avatar evaluation are associated with both avatar selection outcomes and individual personality differences. While these results do not permit direct inferences about specific psychological processes, they provide an insight into the neural correlates of early-stage avatar preference formation in human-computer interaction.

Author(s): Bolden, Khalima Alicia | Advisor(s): Brown, Gregory G | Abstract: The neural substrate of interactions of working memory (WM) with socio-emotional processing is poorly understood in schizophrenia. This study builds on published papers using a delayed match to sample design to study the interaction of WM load with type of distracter (socially relevant faces vs. socially irrelevant geometric designs [FvG]) presented briefly during the WM maintenance period. Based on previously published findings, we hypothesize: (1) The FvG difference in brain activity in the dorsolateral prefrontal cortex (DLPFC) in the task maintenance period will be largest at the highest WM load. (2) Among schizophrenia/ schizoaffective patients and healthy controls the magnitude of the face vs. geometric design (FvG) contrast in brain activity in the amygdala during the task maintenance period will follow a quadratic pattern across WM load when averaged over face type. (3) Among schizophrenia patients, the magnitude of the FvG contrast in brain activity in the amygdala and DLPFC at the greatest WM load will be correlated with negative symptoms. Individuals between the ages of 18-55 diagnosed with schizophrenia/schizoaffective disorder (N = 12) and non-psychiatric controls (N = 20) matched with the patients on age, gender, paternal education and paternal socioeconomic status underwent structural and functional magnetic resonance imaging (fMRI). To assess the effect of implicit socioemotional modulation on brain activity during WM, the effect of facial distraction on brain activation was assessed for WM of pseudowords at three syllable loads (1, 2, and 3) across several face valence types and contrasted with the effect of a geometric distracter. Results: Although patients performed significantly above chance, they were less accurate than controls with no difference in response latency. When the FvG contrast was tested for response latency, we observed a significant quadratic effect of WM load in healthy controls (F[1,19] = 6.108, p = .023, η2 = .24) but a linear effect among patients (F[1,10] = 4.012, p = .073, η2 = .29). Similar patterns were found for response accuracy but were not statistically significant. With regard to neural activity, we found a significant bilateral linear trend of percent signal change on WM load for the FvG contrast in the DLPFC (F[1,19] = 5.818, p = .026, MSE = .077, η2 = .23). among controls, with brain activation to faces greater than activation to designs only at the highest WM load. In the amygdala we observed a significant bilateral quadratic effect of percent signal change on WM load for the FvG contrast in the control group (F[1,19] = 10.423, p = .004, MSE = .121, η2 = .35). We observed a significant difference in neural activation patterns in patients compared to controls in the DLPFC and the amygdala. Specifically, in patients, we observed a quadratic instead of a linear trend in the DLPFC but only in the right hemisphere (hemisphere x quadratic: F[1,11] = 9.362, p = .011, η2 = .40). In the amygdala, the patients displayed a quadratic trend also only in the right hemisphere, (hemisphere x quadratic: F[1,11] = 10.442 p = .008, η2 = .49). In neither controls nor patients did individual differences in the quadratic effect of brain activity in the amygdala correlate with the quadratic effect in response time or accuracy. Although the correlation between the magnitude of the quadratic trend in the right amygdala at the highest WM load with general psychopathology was moderately large in patients, neither this effect nor any other brain activation effects were significantly correlated with psychopathology. Confirming hypothesis one, controls showed the largest difference in brain activity of the FvG contrast in the DLPFC during the maintenance period at the highest WM load. However, in patients we saw significantly decreased percent signal change in DLPFC at the highest WM load on the FvG contrast in the maintenance period. For hypothesis two we observed a quadratic pattern of WM load on the FvG contrast in the maintenance period for both controls and patients, although this effect was only present in the right hemisphere of patients. Furthermore, contrary to hypothesis 3 we did not observe significant correlations between symptom severity and the magnitude of the FvG contrast in brain activity in the amygdala and DLPFC at the greatest WM load. These results suggest a separate process of social-discrimination is taking place in controls. However, this process appears to be impaired in individuals with schizophrenia. This disruption may be due to poor integration of different brain areas and interhemispheric communication.

Transformability, generalizability, but limited diffusibility: Comparing global vs. task-specific language representations in deep neural networks

Understanding how and where in the brain sentence-level meaning is constructed from words presents a major scientific challenge. Recent advances have begun to explain brain activation elicited by sentences using vector models of word meaning derived from patterns of word co-occurrence in text corpora. These studies have helped map out semantic representation across a distributed brain network spanning temporal, parietal, and frontal cortex. However, it remains unclear whether activation patterns within regions reflect unified representations of sentence-level meaning, as opposed to superpositions of context-independent component words. This is because models have typically represented sentences as "bags-of-words" that neglect sentence-level structure. To address this issue, we interrogated fMRI activation elicited as 240 sentences were read by 14 participants (9 female, 5 male), using sentences encoded by a recurrent deep artificial neural-network trained on a sentence inference task (InferSent). Recurrent connections and nonlinear filters enable InferSent to transform sequences of word vectors into unified "propositional" sentence representations suitable for evaluating intersentence entailment relations. Using voxelwise encoding modeling, we demonstrate that InferSent predicts elements of fMRI activation that cannot be predicted by bag-of-words models and sentence models using grammatical rules to assemble word vectors. This effect occurs throughout a distributed network, which suggests that propositional sentence-level meaning is represented within and across multiple cortical regions rather than at any single site. In follow-up analyses, we place results in the context of other deep network approaches (ELMo and BERT) and estimate the degree of unpredicted neural signal using an "experiential" semantic model and cross-participant encoding.SIGNIFICANCE STATEMENT A modern-day scientific challenge is to understand how the human brain transforms word sequences into representations of sentence meaning. A recent approach, emerging from advances in functional neuroimaging, big data, and machine learning, is to computationally model meaning, and use models to predict brain activity. Such models have helped map a cortical semantic information-processing network. However, how unified sentence-level information, as opposed to word-level units, is represented throughout this network remains unclear. This is because models have typically represented sentences as unordered "bags-of-words." Using a deep artificial neural network that recurrently and nonlinearly combines word representations into unified propositional sentence representations, we provide evidence that sentence-level information is encoded throughout a cortical network, rather than in a single region.

When people encounter problems, they typically go through a period of puzzlement and require inspiration from other aspects of life to solve these problems. Friedrich Kekule, for example, discovered the molecular structure of benzene after dreaming of two serpents caught in each other's tails, which formed a circle on a distinctive ring. Archimedes derived insight into the relationship between weight and volume when he noticed displaced water in a bathtub. Numerous major inventions in history have been based on bionic imitation, in which biological systems are applied to the creation of artificial devices and machines. In most previous studies, creativity has been investigated primarily through divergent thinking tasks and insightful problem solving. For example, by comparing the alternative uses task and intelligence-related tasks, different patterns of electrophysiological brain activity and a stronger level of anterior prefrontal brain activation were observed. Although these findings aid the understanding of creativity, whether the progression of scientific invention is identical to the processes discussed in the above-mentioned studies remains unclear. Specifically, the process of prototype position effect of insight in scientific innovation has not been directly addressed. To explore the neural mechanism for prototype position effect of insight in scientific innovation, we selected 40 novel scientific innovation problems (to which scientists have undergone actually but college students did not know the answers) as the material and used fMRI technique for the experiments. Zhang Qinglin proposed that in reality, insight should be caused by "prototypal events activation"; it suggests that the key to solving insight problems is to activate the "prototypal event" correctly and gain "key heuristic information" under experimental conditions.Based on Zhang's hypothesis, this study was designed to explore the neural mechanism governing the act of catching inspiration in scientific innovation. In order to test these hypotheses, we adopted "five to five" paradigm. In this study, the prototype position contains former-problem (problem guide) and former-prototype. The participants were divided into two groupes. First groupes, the participants were asked to investigate 40 post-prototype inventional problems. Second groupes were opposite, the participants were asked to study 40 former-problem inventional problems. The instruments used in the experiments were 40 inventional problems and SPSS 17.0 was used for the statistical analyses. Behavior date showed the mean inventional problem-solving score for former-problem was higher than for former-prototype. And our fMRI data showed that the left middle temporal gyrus、left middle frontal gyrus were significantly activated when college students successfully caught inspiration under former-problem condition; left cingulate gyrus and left middle frontal gyrus were significantly activated when college students successfully solve the problem in the case of former-prototype condition. To the best of our knowledge, this work is the first fMRI study to investigate the brain activation of critical cognitive processes (automatic activation for catching inspiration) behind scientific innovation. Moreover, the real-life scientific innovations used in the experiments have higher ecological validity than the tasks (riddles, remote association tasks, and so on) used in previous studies. That is, the fMRI results can provide valuable insight into the neural basis of catching inspiration.

A single-nucleotide polymorphism at the LHPP gene (rs35936514) has been reported in genome-wide association studies to be associated with major depressive disorder (MDD). However, the neural system effects of rs35936514 that mediate the association are unknown. The present work explores whether the LHPP rs35936514 polymorphism moderates brain regional activity in MDD. A total of 160 subjects were studied: a CC group homozygous for the C allele (23 individuals with MDD and 57 controls) and a T-carrier group carrying the high risk T allele (CT/TT genotypes; 22 MDD and 58 controls). All participants underwent resting-state functional magnetic resonance imaging (rs-fMRI) scanning. Brain activity was assessed using the amplitudes of low-frequency fluctuations (ALFF). MDD patients showed a significant increased ALFF in the left middle temporal gyrus and occipital cortex. The T-carrier group showed increased ALFF in the left superior temporal gyrus. Significant diagnosis × genotype interaction was noted in the bilateral lingual gyri, bilateral dorsal lateral prefrontal cortex (dlPFC), and left medial prefrontal cortex (mPFC) (P < 0.05, corrected). Results demonstrated that MDD patients with LHPP rs35936514 CT/TT genotype may influence the regional brain activity. These findings implicate the effects of the rs35936514 variation on the neural system in MDD.

Understanding spoken words involves complex processes that transform the auditory input into a meaningful interpretation. This effortless transition occurs on millisecond timescales, with remarkable speed and accuracy and without any awareness of the complex computations involved. Here, we reveal the real-time neural dynamics of these processes by collecting data about listeners' brain activity as they hear spoken words. Using novel statistical models of different aspects of the recognition process, we can locate directly which parts of the brain are accessing the stored form and meaning of words and how the competition between different word candidates is resolved neurally in real time. This gives us a uniquely differentiated picture of the neural substrate for the first 500 ms of word recognition.

The problem of visual recognition entails matching the sensory input with the stored knowledge of semantic information. This process involves a feed-forward component where visual input is processed to extract features characterizing different objects. It is also presumed to involve an opposite processing pathway where high-level features propagate back providing prediction on the kind of visual stimulus presented. This top-down pathway appears to be useful particularly for tasks like processing degraded stimuli. The process by which the forward and backward pathways integrate leading to visual perception is still largely unknown. Here, using a deep neural network (DNN) trained to recognize objects as a proxy for hierarchical neural representations (Horikawa & Kamitani, 2015), we demonstrate that top-down processing pathway attempts to supplement missing visual features. We first trained multivoxel fMRI decoders to predict DNN features of multiple layers for stimulus images. The trained decoders were then used to analyze independent fMRI data collected while viewing pairs of normal and degraded images. Degraded images were created by blurring original images using averaging filters, and by binarizing slightly blurred images using thresholding. We found that decoded features from fMRI responses to degraded images were more correlated to DNN features calculated from the original images than from the degraded (presented) ones. This was especially salient in the lower level DNN representations. We also found that the task of categorizing the visual stimuli increased the correlation difference especially in higher visual areas. These results suggest the operation of the top-down pathway as it attempts to supplement the missing information in degraded images. The effect of the categorizing task may indicate how giving a prior guides the exploration efforts towards successful perception. This DNN-based brain decoding approach may reveal the interactions between the bottom-up and the top-down pathways, providing empirical evidence for existing and novel theoretical models. Meeting abstract presented at VSS 2017

BackgroundIndividuals with Autism Spectrum Disorder (ASD) consistently exhibit difficulties in comprehending figurative language. While prior neuroimaging studies have identified discrepancies in brain activation between ASD individuals and neurotypical controls (NCs) during such tasks, the specific and consistent neural patterns underlying these deficits remain unclear.MethodsWe conducted an activation likelihood estimation (ALE) meta-analysis to identify consistent brain activation patterns associated with figurative language comprehension in ASD. A systematic literature search was performed across five databases (PubMed, Embase, Web of Science, CNKI, Wanfang) up to January 31, 2025. Six functional magnetic resonance imaging (fMRI) studies, comprising 95 individuals with ASD and 98 NCs, met the inclusion criteria.ResultsThe analysis revealed that during figurative language comprehension, both ASD and NCs consistently activated a core network involving the bilateral superior temporal gyrus (STG), transverse temporal gyrus, and right insula. A conjunction analysis confirmed the stability of these shared regions. Crucially, the meta-analysis of group differences (NCs > ASD) identified a significant cluster of hypoactivation in the ASD group within the left STG and middle temporal gyrus (MTG). No significant hyperactivation was found in the ASD group compared to NCs.ConclusionThis meta-analysis demonstrates that while individuals with ASD recruit typical language networks during figurative language comprehension, they exhibit a consistent pattern of reduced neural recruitment in the left STG and MTG. This hypoactivation may reflect a dual deficit encompassing reduced efficiency in semantic access/integration and impaired socio-linguistic integration, providing a neurobiological substrate for the pragmatic language difficulties characteristic of ASD.Systematic review registrationhttps://www.crd.york.ac.uk/PROSPERO/view/CRD42023467185, CRD42023467185.

Comparing neural activity during autobiographical memory retrieval between younger and older adults: An ALE meta-analysis

Conventional blood oxygenation level-dependent (BOLD) based functional magnetic resonance imaging (fMRI) is accompanied by substantial acoustic gradient noise. This noise can influence the performance as well as neuronal activations. Conventional fMRI typically has a pulsed noise component, which is a particularly efficient auditory stimulus. We investigated whether the elimination of this pulsed noise component in a recent modification of continuous-sound fMRI modifies neuronal activations in a cognitively demanding non-auditory working memory task. Sixteen normal subjects performed a letter variant n-back task. Brain activity and psychomotor performance was examined during fMRI with continuous-sound fMRI and conventional fMRI. We found greater BOLD responses in bilateral medial frontal gyrus, left middle frontal gyrus, left middle temporal gyrus, left hippocampus, right superior frontal gyrus, right precuneus and right cingulate gyrus with continuous-sound compared to conventional fMRI. Conversely, BOLD responses were greater in bilateral cingulate gyrus, left middle and superior frontal gyrus and right lingual gyrus with conventional compared to continuous-sound fMRI. There were no differences in psychomotor performance between both scanning protocols. Although behavioral performance was not affected, acoustic gradient noise interferes with neuronal activations in non-auditory cognitive tasks and represents a putative systematic confound.

This study aimed to assess how a Zentangle intervention influences cognitive focus, emotional well-being, stress levels, and neural activity patterns across brain regions and frequency bands. A cohort of 30 healthy adults, all without prior Zentangle experience, participated in this study. Electroencephalography (EEG) was used to measure their brain activity, and self-reported data were collected through questionnaires to assess subjects' concentration levels, emotional calm, and stress and anxiety. Participants reported enhanced cognitive focus and emotional well-being, evidenced by increased self-reported concentration and emotional calmness, and reduced stress and anxiety levels during the intervention. EEG analyses revealed notable changes in neural activity patterns, including decreased delta power and increased theta, alpha, beta, and gamma bands. Functional connectivity analysis also highlighted alterations in the brain's functional connectivity, suggesting potential effects on neural communication and information processing. This study provides compelling evidence of Zentangle's impact on EEG data, aligning it with equanimity and tranquility consistent with previous mindfulness research. These findings underscore Zentangle as an effective mindfulness practice, potentially enhancing cognitive focus and emotional well-being, and emerging as a valuable intervention for improving mental health and overall well-being.