The lexical semantics of adjective-noun phrases in the human brain.

As a person reads, the brain performs complex operations to create higher order semantic representations from individual words. While these steps are effortless for competent readers, we are only beginning to understand how the brain performs these actions. Here, we explore lexical semantics using magnetoencephalography (MEG) recordings of people reading adjective-noun phrases presented one word at a time. We track the neural representation of single word representations over time, through different brain regions. Our results reveal two novel findings: (a) a neural representation of the adjective is present during noun presentation, but this representation is different from that observed during adjective presentation and (b) the neural representation of adjective semantics observed during adjective reading is reactivated after phrase reading, with remarkable consistency. We also note that while the semantic representation of the adjective during the reading of the adjective is very distributed, the later representations are concentrated largely to temporal and frontal areas previously associated with composition. Taken together, these results paint a picture of information flow in the brain as phrases are read and understood.

Exploring commonalities across participants in the neural representation of objects.

The question of whether the neural encodings of objects are similar across different people is one of the key questions in cognitive neuroscience. This article examines the commonalities in the internal representation of objects, as measured with fMRI, across individuals in two complementary ways. First, we examine the commonalities in the internal representation of objects across people at the level of interobject distances, derived from whole brain fMRI data, and second, at the level of spatially localized anatomical brain regions that contain sufficient information for identification of object categories, without making the assumption that their voxel patterns are spatially matched in a common space. We examine the commonalities in internal representation of objects on 3T fMRI data collected while participants viewed line drawings depicting various tools and dwellings. This exploratory study revealed the extent to which the representation of individual concepts, and their mutual similarity, is shared across participants.

Erratum: Gender bias, social bias, and representation in Bollywood and Hollywood.

[This corrects the article DOI: 10.1016/j.patter.2022.100442.][This corrects the article DOI: 10.1016/j.patter.2021.100409.].

Combining computational controls with natural text reveals aspects of meaning composition.

To study a core component of human intelligence-our ability to combine the meaning of words-neuroscientists have looked to linguistics. However, linguistic theories are insufficient to account for all brain responses reflecting linguistic composition. In contrast, we adopt a data-driven approach to study the composed meaning of words beyond their individual meaning, which we term 'supra-word meaning'. We construct a computational representation for supra-word meaning and study its brain basis through brain recordings from two complementary imaging modalities. Using functional magnetic resonance imaging, we reveal that hubs that are thought to process lexical meaning also maintain supra-word meaning, suggesting a common substrate for lexical and combinatorial semantics. Surprisingly, we cannot detect supra-word meaning in magnetoencephalography, which suggests that composed meaning might be maintained through a different neural mechanism than the synchronized firing of pyramidal cells. This sensitivity difference has implications for past neuroimaging results and future wearable neurotechnology.

Gender bias, social bias, and representation: 70 years of B H ollywood.

We use a suite of cutting-edge natural language processing methods to quantify and characterize societal and gender biases in popular movie content. Our data set consists of English subtitles of popular movies from Bollywood-the Mumbai film industry-spanning 7 decades (700 movies). In addition, we include movies from Hollywood and movies nominated for the Academy Awards for contrastive purposes. Our findings indicate that while the overall portrayal of women has improved over time in popular movie dialogues from both Bollywood and Hollywood, modern films still exhibit considerable gender bias and are yet to achieve equal representation among genders. We also observe a strong bias favoring fair skin color in Bollywood content that occurred consistently across all time periods we considered. While our geographic representation analysis indicates improved inclusion over time for several Indian states, it also reveals a long-standing under-representation of many northeastern Indian states.

Erratum: Gender bias, social bias, and representation: 70 years of B H ollywood.

[This corrects the article DOI: 10.1016/j.patter.2021.100409.].

Commonality of neural representations of words and pictures.

In this work we explore whether the patterns of brain activity associated with thinking about concrete objects are dependent on stimulus presentation format, whether an object is referred to by a written or pictorial form. Multi-voxel pattern analysis methods were applied to brain imaging (fMRI) data to identify the item category associated with brief viewings of each of 10 words (naming 5 tools and 5 dwellings) and, separately, with brief viewings of each of 10 pictures (line drawings) of the objects named by the words. These methods were able to identify the category of the picture the participant was viewing, based on neural activation patterns observed during word-viewing, and identify the category of the word the participant was viewing, based on neural activation patterns observed during picture-viewing, using data from only that participant or only from other participants. These results provide an empirical demonstration of object category identification across stimulus formats and across participants. In addition, we were able to identify the category of the word that the participant was viewing based on the patterns of neural activation generated during word-viewing by that participant or by all other participants. Similarly, we were able to identify with even higher accuracy the category of the picture the participant was viewing, based on the patterns of neural activation demonstrated during picture-viewing by that participant or by all other participants. The brain locations that were important for category identification were similar across participants and were distributed throughout the cortex where various object properties might be neurally represented. These findings indicate consistent triggering of semantic representations using different stimulus formats and suggest the presence of stable, distributed, and identifiable neural states that are common to pictorial and verbal input referring to object categories.

Modeling fMRI data generated by overlapping cognitive processes with unknown onsets using Hidden Process Models.

We present a new method for modeling fMRI time series data called Hidden Process Models (HPMs). Like several earlier models for fMRI analysis, Hidden Process Models assume that the observed data is generated by a sequence of underlying mental processes that may be triggered by stimuli. HPMs go beyond these earlier models by allowing for processes whose timing may be unknown, and that might not be directly tied to specific stimuli. HPMs provide a principled, probabilistic framework for simultaneously learning the contribution of each process to the observed data, as well as the timing and identities of each instantiated process. They also provide a framework for evaluating and selecting among competing models that assume different numbers and types of underlying mental processes. We describe the HPM framework and its learning and inference algorithms, and present experimental results demonstrating its use on simulated and real fMRI data. Our experiments compare several models of the data using cross-validated data log-likelihood in an fMRI study involving overlapping mental processes whose timings are not fully known.

Turbo-SMT: Parallel Coupled Sparse Matrix-Tensor Factorizations and Applications.

How can we correlate the neural activity in the human brain as it responds to typed words, with properties of these terms (like 'edible', 'fits in hand')? In short, we want to find latent variables, that jointly explain both the brain activity, as well as the behavioral responses. This is one of many settings of the Coupled Matrix-Tensor Factorization (CMTF) problem. Can we enhance any CMTF solver, so that it can operate on potentially very large datasets that may not fit in main memory? We introduce Turbo-SMT, a meta-method capable of doing exactly that: it boosts the performance of any CMTF algorithm, produces sparse and interpretable solutions, and parallelizes any CMTF algorithm, producing sparse and interpretable solutions (up to 65 fold). Additionally, we improve upon ALS, the work-horse algorithm for CMTF, with respect to efficiency and robustness to missing values. We apply Turbo-SMT to BrainQ, a dataset consisting of a (nouns, brain voxels, human subjects) tensor and a (nouns, properties) matrix, with coupling along the nouns dimension. Turbo-SMT is able to find meaningful latent variables, as well as to predict brain activity with competitive accuracy. Finally, we demonstrate the generality of Turbo-SMT, by applying it on a Facebook dataset (users, 'friends', wall-postings); there, Turbo-SMT spots spammer-like anomalies.

Interpretable Semantic Vectors from a Joint Model of Brain- and Text-Based Meaning.

Vector space models (VSMs) represent word meanings as points in a high dimensional space. VSMs are typically created using a large text corpora, and so represent word semantics as observed in text. We present a new algorithm (JNNSE) that can incorporate a measure of semantics not previously used to create VSMs: brain activation data recorded while people read words. The resulting model takes advantage of the complementary strengths and weaknesses of corpus and brain activation data to give a more complete representation of semantics. Evaluations show that the model 1) matches a behavioral measure of semantics more closely, 2) can be used to predict corpus data for unseen words and 3) has predictive power that generalizes across brain imaging technologies and across subjects. We believe that the model is thus a more faithful representation of mental vocabularies.

Good-Enough Brain Model: Challenges, Algorithms, and Discoveries in Multisubject Experiments.

Given a simple noun such as apple, and a question such as "Is it edible?," what processes take place in the human brain? More specifically, given the stimulus, what are the interactions between (groups of) neurons (also known as functional connectivity) and how can we automatically infer those interactions, given measurements of the brain activity? Furthermore, how does this connectivity differ across different human subjects? In this work, we show that this problem, even though originating from the field of neuroscience, can benefit from big data techniques; we present a simple, novel good-enough brain model, or GeBM in short, and a novel algorithm Sparse-SysId, which are able to effectively model the dynamics of the neuron interactions and infer the functional connectivity. Moreover, GeBM is able to simulate basic psychological phenomena such as habituation and priming (whose definition we provide in the main text). We evaluate GeBM by using real brain data. GeBM produces brain activity patterns that are strikingly similar to the real ones, where the inferred functional connectivity is able to provide neuroscientific insights toward a better understanding of the way that neurons interact with each other, as well as detect regularities and outliers in multisubject brain activity measurements.

Identifying autism from neural representations of social interactions: neurocognitive markers of autism.

Autism is a psychiatric/neurological condition in which alterations in social interaction (among other symptoms) are diagnosed by behavioral psychiatric methods. The main goal of this study was to determine how the neural representations and meanings of social concepts (such as to insult) are altered in autism. A second goal was to determine whether these alterations can serve as neurocognitive markers of autism. The approach is based on previous advances in fMRI analysis methods that permit (a) the identification of a concept, such as the thought of a physical object, from its fMRI pattern, and (b) the ability to assess the semantic content of a concept from its fMRI pattern. These factor analysis and machine learning methods were applied to the fMRI activation patterns of 17 adults with high-functioning autism and matched controls, scanned while thinking about 16 social interactions. One prominent neural representation factor that emerged (manifested mainly in posterior midline regions) was related to self-representation, but this factor was present only for the control participants, and was near-absent in the autism group. Moreover, machine learning algorithms classified individuals as autistic or control with 97% accuracy from their fMRI neurocognitive markers. The findings suggest that psychiatric alterations of thought can begin to be biologically understood by assessing the form and content of the altered thought's underlying brain activation patterns.

Turbo-SMT: Accelerating Coupled Sparse Matrix-Tensor Factorizations by 200×.

How can we correlate the neural activity in the human brain as it responds to typed words, with properties of these terms (like 'edible', 'fits in hand')? In short, we want to find latent variables, that jointly explain both the brain activity, as well as the behavioral responses. This is one of many settings of the Coupled Matrix-Tensor Factorization (CMTF) problem. Can we accelerate any CMTF solver, so that it runs within a few minutes instead of tens of hours to a day, while maintaining good accuracy? We introduce TURBO-SMT, a meta-method capable of doing exactly that: it boosts the performance of any CMTF algorithm, by up to 200×, along with an up to 65 fold increase in sparsity, with comparable accuracy to the baseline. We apply TURBO-SMT to BRAINQ, a dataset consisting of a (nouns, brain voxels, human subjects) tensor and a (nouns, properties) matrix, with coupling along the nouns dimension. TURBO-SMT is able to find meaningful latent variables, as well as to predict brain activity with competitive accuracy.

Identifying bilingual semantic neural representations across languages.

The goal of the study was to identify the neural representation of a noun's meaning in one language based on the neural representation of that same noun in another language. Machine learning methods were used to train classifiers to identify which individual noun bilingual participants were thinking about in one language based solely on their brain activation in the other language. The study shows reliable (p<.05) pattern-based classification accuracies for the classification of brain activity for nouns across languages. It also shows that the stable voxels used to classify the brain activation were located in areas associated with encoding information about semantic dimensions of the words in the study. The identification of the semantic trace of individual nouns from the pattern of cortical activity demonstrates the existence of a multi-voxel pattern of activation across the cortex for a single noun common to both languages in bilinguals.

A neurosemantic theory of concrete noun representation based on the underlying brain codes.

This article describes the discovery of a set of biologically-driven semantic dimensions underlying the neural representation of concrete nouns, and then demonstrates how a resulting theory of noun representation can be used to identify simple thoughts through their fMRI patterns. We use factor analysis of fMRI brain imaging data to reveal the biological representation of individual concrete nouns like apple, in the absence of any pictorial stimuli. From this analysis emerge three main semantic factors underpinning the neural representation of nouns naming physical objects, which we label manipulation, shelter, and eating. Each factor is neurally represented in 3-4 different brain locations that correspond to a cortical network that co-activates in non-linguistic tasks, such as tool use pantomime for the manipulation factor. Several converging methods, such as the use of behavioral ratings of word meaning and text corpus characteristics, provide independent evidence of the centrality of these factors to the representations. The factors are then used with machine learning classifier techniques to show that the fMRI-measured brain representation of an individual concrete noun like apple can be identified with good accuracy from among 60 candidate words, using only the fMRI activity in the 16 locations associated with these factors. To further demonstrate the generativity of the proposed account, a theory-based model is developed to predict the brain activation patterns for words to which the algorithm has not been previously exposed. The methods, findings, and theory constitute a new approach of using brain activity for understanding how object concepts are represented in the mind.

Using FMRI brain activation to identify cognitive states associated with perception of tools and dwellings.

Previous studies have succeeded in identifying the cognitive state corresponding to the perception of a set of depicted categories, such as tools, by analyzing the accompanying pattern of brain activity, measured with fMRI. The current research focused on identifying the cognitive state associated with a 4s viewing of an individual line drawing (1 of 10 familiar objects, 5 tools and 5 dwellings, such as a hammer or a castle). Here we demonstrate the ability to reliably (1) identify which of the 10 drawings a participant was viewing, based on that participant's characteristic whole-brain neural activation patterns, excluding visual areas; (2) identify the category of the object with even higher accuracy, based on that participant's activation; and (3) identify, for the first time, both individual objects and the category of the object the participant was viewing, based only on other participants' activation patterns. The voxels important for category identification were located similarly across participants, and distributed throughout the cortex, focused in ventral temporal perceptual areas but also including more frontal association areas (and somewhat left-lateralized). These findings indicate the presence of stable, distributed, communal, and identifiable neural states corresponding to object concepts.

Predicting human brain activity associated with the meanings of nouns.

The question of how the human brain represents conceptual knowledge has been debated in many scientific fields. Brain imaging studies have shown that different spatial patterns of neural activation are associated with thinking about different semantic categories of pictures and words (for example, tools, buildings, and animals). We present a computational model that predicts the functional magnetic resonance imaging (fMRI) neural activation associated with words for which fMRI data are not yet available. This model is trained with a combination of data from a trillion-word text corpus and observed fMRI data associated with viewing several dozen concrete nouns. Once trained, the model predicts fMRI activation for thousands of other concrete nouns in the text corpus, with highly significant accuracies over the 60 nouns for which we currently have fMRI data.

Modeling the fMRI signal via Hierarchical Clustered Hidden Process Models.

Machine Learning techniques have been used quite widely for the task of predicting cognitive processes from fMRI data. However, these models do not describe well the fMRI signal when it is generated by multiple cognitive processes that are simultaneously active. In this paper we consider the problem of accurately modeling the fMRI signal of a human subject who is performing a task involving multiple concurrent cognitive processes. We present a Hierarchical Clustering extension of Hidden Process Models which, by taking advantage of automatically discovered similarities in the activation among neighboring voxels, achieves significantly better performance than standard generative models in terms of Average Log Likelihood.