Stacked regressions and structured variance partitioning for interpretable brain maps.

Relating brain activity associated with a complex stimulus to different properties of that stimulus is a powerful approach for constructing functional brain maps. However, when stimuli are naturalistic, their properties are often correlated (e.g., visual and semantic features of natural images, or different layers of a convolutional neural network that are used as features of images). Correlated properties can act as confounders for each other and complicate the interpretability of brain maps, and can impact the robustness of statistical estimators. Here, we present an approach for brain mapping based on two proposed methods: stacking different encoding models and structured variance partitioning. Our stacking algorithm combines encoding models that each uses as input a feature space that describes a different stimulus attribute. The algorithm learns to predict the activity of a voxel as a linear combination of the outputs of different encoding models. We show that the resulting combined model can predict held-out brain activity better or at least as well as the individual encoding models. Further, the weights of the linear combination are readily interpretable; they show the importance of each feature space for predicting a voxel. We then build on our stacking models to introduce structured variance partitioning, a new type of variance partitioning that takes into account the known relationships between features. Our approach constrains the size of the hypothesis space and allows us to ask targeted questions about the similarity between feature spaces and brain regions even in the presence of correlations between the feature spaces. We validate our approach in simulation, showcase its brain mapping potential on fMRI data, and release a Python package. Our methods can be useful for researchers interested in aligning brain activity with different layers of a neural network, or with other types of correlated feature spaces.

Improved quantitative parameter estimation for prostate T2 relaxometry using convolutional neural networks.

OBJECTIVE: Quantitative parameter mapping conventionally relies on curve fitting techniques to estimate parameters from magnetic resonance image series. This study compares conventional curve fitting techniques to methods using neural networks (NN) for measuring T2 in the prostate. MATERIALS AND METHODS: Large physics-based synthetic datasets simulating T2 mapping acquisitions were generated for training NNs and for quantitative performance comparisons. Four combinations of different NN architectures and training corpora were implemented and compared with four different curve fitting strategies. All methods were compared quantitatively using synthetic data with known ground truth, and further compared on in vivo test data, with and without noise augmentation, to evaluate feasibility and noise robustness. RESULTS: In the evaluation on synthetic data, a convolutional neural network (CNN), trained in a supervised fashion using synthetic data generated from naturalistic images, showed the highest overall accuracy and precision amongst the methods. On in vivo data, this best performing method produced low-noise T2 maps and showed the least deterioration with increasing input noise levels. DISCUSSION: This study showed that a CNN, trained with synthetic data in a supervised manner, may provide superior T2 estimation performance compared to conventional curve fitting, especially in low signal-to-noise regions.

Temporal asymmetry of neural representations predicts memory decisions.

A stimulus can be familiar for multiple reasons. It might have been recently encountered, or is similar to recent experience, or is similar to 'typical' experience. Understanding how the brain translates these sources of similarity into memory decisions is a fundamental, but challenging goal. Here, using fMRI, we computed neural similarity between a current stimulus and events from different temporal windows in the past and future (from seconds to days). We show that trial-by-trial memory decisions (is this stimulus 'old'?) were predicted by the difference in similarity to past vs. future events (temporal asymmetry). This relationship was (i) evident in lateral parietal and occipitotemporal cortices, (ii) strongest when considering events from the recent past (minutes ago), and (iii) most pronounced when veridical (true) memories were weak. These findings suggest a new perspective in which the brain supports memory decisions by comparing what actually occurred to what is likely to occur.

Modeling short visual events through the BOLD moments video fMRI dataset and metadata.

Studying the neural basis of human dynamic visual perception requires extensive experimental data to evaluate the large swathes of functionally diverse brain neural networks driven by perceiving visual events. Here, we introduce the BOLD Moments Dataset (BMD), a repository of whole-brain fMRI responses to over 1000 short (3 s) naturalistic video clips of visual events across ten human subjects. We use the videos' extensive metadata to show how the brain represents word- and sentence-level descriptions of visual events and identify correlates of video memorability scores extending into the parietal cortex. Furthermore, we reveal a match in hierarchical processing between cortical regions of interest and video-computable deep neural networks, and we showcase that BMD successfully captures temporal dynamics of visual events at second resolution. With its rich metadata, BMD offers new perspectives and accelerates research on the human brain basis of visual event perception.

Natural scenes reveal diverse representations of 2D and 3D body pose in the human brain.

Human pose, defined as the spatial relationships between body parts, carries instrumental information supporting the understanding of motion and action of a person. A substantial body of previous work has identified cortical areas responsive to images of bodies and different body parts. However, the neural basis underlying the visual perception of body part relationships has received less attention. To broaden our understanding of body perception, we analyzed high-resolution fMRI responses to a wide range of poses from over 4,000 complex natural scenes. Using ground-truth annotations and an application of three-dimensional (3D) pose reconstruction algorithms, we compared similarity patterns of cortical activity with similarity patterns built from human pose models with different levels of depth availability and viewpoint dependency. Targeting the challenge of explaining variance in complex natural image responses with interpretable models, we achieved statistically significant correlations between pose models and cortical activity patterns (though performance levels are substantially lower than the noise ceiling). We found that the 3D view-independent pose model, compared with two-dimensional models, better captures the activation from distinct cortical areas, including the right posterior superior temporal sulcus (pSTS). These areas, together with other pose-selective regions in the LOTC, form a broader, distributed cortical network with greater view-tolerance in more anterior patches. We interpret these findings in light of the computational complexity of natural body images, the wide range of visual tasks supported by pose structures, and possible shared principles for view-invariant processing between articulated objects and ordinary, rigid objects.

Disentangling signal and noise in neural responses through generative modeling.

Measurements of neural responses to identically repeated experimental events often exhibit large amounts of variability. This noise is distinct from signal, operationally defined as the average expected response across repeated trials for each given event. Accurately distinguishing signal from noise is important, as each is a target that is worthy of study (many believe noise reflects important aspects of brain function) and it is important not to confuse one for the other. Here, we introduce a principled modeling approach in which response measurements are explicitly modeled as the sum of samples from multivariate signal and noise distributions. In our proposed method-termed Generative Modeling of Signal and Noise (GSN)-the signal distribution is estimated by subtracting the estimated noise distribution from the estimated data distribution. We validate GSN using ground-truth simulations and demonstrate the application of GSN to empirical fMRI data. In doing so, we illustrate a simple consequence of GSN: by disentangling signal and noise components in neural responses, GSN denoises principal components analysis and improves estimates of dimensionality. We end by discussing other situations that may benefit from GSN's characterization of signal and noise, such as estimation of noise ceilings for computational models of neural activity. A code toolbox for GSN is provided with both MATLAB and Python implementations.

Driving and suppressing the human language network using large language models.

Transformer models such as GPT generate human-like language and are predictive of human brain responses to language. Here, using functional-MRI-measured brain responses to 1,000 diverse sentences, we first show that a GPT-based encoding model can predict the magnitude of the brain response associated with each sentence. We then use the model to identify new sentences that are predicted to drive or suppress responses in the human language network. We show that these model-selected novel sentences indeed strongly drive and suppress the activity of human language areas in new individuals. A systematic analysis of the model-selected sentences reveals that surprisal and well-formedness of linguistic input are key determinants of response strength in the language network. These results establish the ability of neural network models to not only mimic human language but also non-invasively control neural activity in higher-level cortical areas, such as the language network.

Social cognitive regions of human association cortex are selectively connected to the amygdala.

Reasoning about someone's thoughts and intentions - i.e., forming a theory of mind - is an important aspect of social cognition that relies on association areas of the brain that have expanded disproportionately in the human lineage. We recently showed that these association zones comprise parallel distributed networks that, despite occupying adjacent and interdigitated regions, serve dissociable functions. One network is selectively recruited by theory of mind processes. What circuit properties differentiate these parallel networks? Here, we show that social cognitive association areas are intrinsically and selectively connected to regions of the anterior medial temporal lobe that are implicated in emotional learning and social behaviors, including the amygdala at or near the basolateral complex and medial nucleus. The results suggest that social cognitive functions emerge through coordinated activity between amygdala circuits and a distributed association network, and indicate the medial nucleus may play an important role in social cognition in humans.

Expansion of a frontostriatal salience network in individuals with depression.

Hundreds of neuroimaging studies spanning two decades have revealed differences in brain structure and functional connectivity in depression, but with modest effect sizes, complicating efforts to derive mechanistic pathophysiologic insights or develop biomarkers. 1 Furthermore, although depression is a fundamentally episodic condition, few neuroimaging studies have taken a longitudinal approach, which is critical for understanding cause and effect and delineating mechanisms that drive mood state transitions over time. The emerging field of precision functional mapping using densely-sampled longitudinal neuroimaging data has revealed unexpected, functionally meaningful individual differences in brain network topology in healthy individuals, 2-5 but these approaches have never been applied to individuals with depression. Here, using precision functional mapping techniques and 11 datasets comprising n=187 repeatedly sampled individuals and >21,000 minutes of fMRI data, we show that the frontostriatal salience network is expanded two-fold in most individuals with depression. This effect was replicable in multiple samples, including large-scale, group-average data (N=1,231 subjects), and caused primarily by network border shifts affecting specific functional systems, with three distinct modes of encroachment occurring in different individuals. Salience network expansion was unexpectedly stable over time, unaffected by changes in mood state, and detectable in children before the subsequent onset of depressive symptoms in adolescence. Longitudinal analyses of individuals scanned up to 62 times over 1.5 years identified connectivity changes in specific frontostriatal circuits that tracked fluctuations in specific symptom domains and predicted future anhedonia symptoms before they emerged. Together, these findings identify a stable trait-like brain network topology that may confer risk for depression and mood-state dependent connectivity changes in frontostriatal circuits that predict the emergence and remission of depressive symptoms over time.

Re-expression of CA1 and entorhinal activity patterns preserves temporal context memory at long timescales.

Converging, cross-species evidence indicates that memory for time is supported by hippocampal area CA1 and entorhinal cortex. However, limited evidence characterizes how these regions preserve temporal memories over long timescales (e.g., months). At long timescales, memoranda may be encountered in multiple temporal contexts, potentially creating interference. Here, using 7T fMRI, we measured CA1 and entorhinal activity patterns as human participants viewed thousands of natural scene images distributed, and repeated, across many months. We show that memory for an image's original temporal context was predicted by the degree to which CA1/entorhinal activity patterns from the first encounter with an image were re-expressed during re-encounters occurring minutes to months later. Critically, temporal memory signals were dissociable from predictors of recognition confidence, which were carried by distinct medial temporal lobe expressions. These findings suggest that CA1 and entorhinal cortex preserve temporal memories across long timescales by coding for and reinstating temporal context information.

Brain-optimized deep neural network models of human visual areas learn non-hierarchical representations.

Deep neural networks (DNNs) optimized for visual tasks learn representations that align layer depth with the hierarchy of visual areas in the primate brain. One interpretation of this finding is that hierarchical representations are necessary to accurately predict brain activity in the primate visual system. To test this interpretation, we optimized DNNs to directly predict brain activity measured with fMRI in human visual areas V1-V4. We trained a single-branch DNN to predict activity in all four visual areas jointly, and a multi-branch DNN to predict each visual area independently. Although it was possible for the multi-branch DNN to learn hierarchical representations, only the single-branch DNN did so. This result shows that hierarchical representations are not necessary to accurately predict human brain activity in V1-V4, and that DNNs that encode brain-like visual representations may differ widely in their architecture, ranging from strict serial hierarchies to multiple independent branches.

Stacked regressions and structured variance partitioning for interpretable brain maps.

Relating brain activity associated with a complex stimulus to different properties of that stimulus is a powerful approach for constructing functional brain maps. However, when stimuli are naturalistic, their properties are often correlated (e.g., visual and semantic features of natural images, or different layers of a convolutional neural network that are used as features of images). Correlated properties can act as confounders for each other and complicate the interpretability of brain maps, and can impact the robustness of statistical estimators. Here, we present an approach for brain mapping based on two proposed methods: stacking different encoding models and structured variance partitioning. Our stacking algorithm combines encoding models that each use as input a feature space that describes a different stimulus attribute. The algorithm learns to predict the activity of a voxel as a linear combination of the outputs of different encoding models. We show that the resulting combined model can predict held-out brain activity better or at least as well as the individual encoding models. Further, the weights of the linear combination are readily interpretable; they show the importance of each feature space for predicting a voxel. We then build on our stacking models to introduce structured variance partitioning, a new type of variance partitioning that takes into account the known relationships between features. Our approach constrains the size of the hypothesis space and allows us to ask targeted questions about the similarity between feature spaces and brain regions even in the presence of correlations between the feature spaces. We validate our approach in simulation, showcase its brain mapping potential on fMRI data, and release a Python package. Our methods can be useful for researchers interested in aligning brain activity with different layers of a neural network, or with other types of correlated feature spaces.

Tasks and their role in visual neuroscience.

Vision is widely used as a model system to gain insights into how sensory inputs are processed and interpreted by the brain. Historically, careful quantification and control of visual stimuli have served as the backbone of visual neuroscience. There has been less emphasis, however, on how an observer's task influences the processing of sensory inputs. Motivated by diverse observations of task-dependent activity in the visual system, we propose a framework for thinking about tasks, their role in sensory processing, and how we might formally incorporate tasks into our models of vision.

Driving and suppressing the human language network using large language models.

Transformer models such as GPT generate human-like language and are highly predictive of human brain responses to language. Here, using fMRI-measured brain responses to 1,000 diverse sentences, we first show that a GPT-based encoding model can predict the magnitude of brain response associated with each sentence. Then, we use the model to identify new sentences that are predicted to drive or suppress responses in the human language network. We show that these model-selected novel sentences indeed strongly drive and suppress activity of human language areas in new individuals. A systematic analysis of the model-selected sentences reveals that surprisal and well-formedness of linguistic input are key determinants of response strength in the language network. These results establish the ability of neural network models to not only mimic human language but also noninvasively control neural activity in higher-level cortical areas, like the language network.

Engaging in word recognition elicits highly specific modulations in visual cortex.

A person's cognitive state determines how their brain responds to visual stimuli. The most common such effect is a response enhancement when stimuli are task relevant and attended rather than ignored. In this fMRI study, we report a surprising twist on such attention effects in the visual word form area (VWFA), a region that plays a key role in reading. We presented participants with strings of letters and visually similar shapes, which were either relevant for a specific task (lexical decision or gap localization) or ignored (during a fixation dot color task). In the VWFA, the enhancement of responses to attended stimuli occurred only for letter strings, whereas non-letter shapes evoked smaller responses when attended than when ignored. The enhancement of VWFA activity was accompanied by strengthened functional connectivity with higher-level language regions. These task-dependent modulations of response magnitude and functional connectivity were specific to the VWFA and absent in the rest of visual cortex. We suggest that language regions send targeted excitatory feedback into the VWFA only when the observer is trying to read. This feedback enables the discrimination of familiar and nonsense words and is distinct from generic effects of visual attention.

Improved Quantitative Parameter Estimation for Prostate T2 Relaxometry using Convolutional Neural Networks.

This work seeks to evaluate multiple methods for quantitative parameter estimation from standard T2 mapping acquisitions in the prostate. The T2 estimation performance of methods based on neural networks (NN) was quantitatively compared to that of conventional curve fitting techniques. Large physics-based synthetic datasets simulating T2 mapping acquisitions were generated for training NNs and for quantitative performance comparisons. Ten combinations of different NN architectures, training strategies, and training corpora were implemented and compared with four different curve fitting strategies. All methods were compared quantitatively using synthetic data with known ground truth, and further compared on in vivo test data, with and without noise augmentation, to evaluate feasibility and noise robustness. In the evaluation on synthetic data, a convolutional neural network (CNN), trained in a supervised fashion using synthetic data generated from naturalistic images, showed the highest overall accuracy and precision amongst all the methods. On in vivo data, this best-performing method produced low-noise T2 maps and showed the least deterioration with increasing input noise levels. This study showed that a CNN, trained with synthetic data in a supervised manner, may provide superior T2 estimation performance compared to conventional curve fitting, especially in low signal-to-noise regions.

Color-biased regions in the ventral visual pathway are food selective.

Color-biased regions have been found between face- and place-selective areas in the ventral visual pathway. To investigate the function of the color-biased regions in a pathway responsible for object recognition, we analyzed the natural scenes dataset (NSD), a large 7T fMRI dataset from 8 participants who each viewed up to 30,000 trials of images of colored natural scenes over more than 30 scanning sessions. In a whole-brain analysis, we correlated the average color saturation of the images with voxel responses, revealing color-biased regions that diverge into two streams, beginning in V4 and extending medially and laterally relative to the fusiform face area in both hemispheres. We drew regions of interest (ROIs) for the two streams and found that the images for each ROI that evoked the largest responses had certain characteristics: they contained food, circular objects, warmer hues, and had higher color saturation. Further analyses showed that food images were the strongest predictor of activity in these regions, implying the existence of medial and lateral ventral food streams (VFSs). We found that color also contributed independently to voxel responses, suggesting that the medial and lateral VFSs use both color and form to represent food. Our findings illustrate how high-resolution datasets such as the NSD can be used to disentangle the multifaceted contributions of many visual features to the neural representations of natural scenes.

Improving the accuracy of single-trial fMRI response estimates using GLMsingle.

Advances in artificial intelligence have inspired a paradigm shift in human neuroscience, yielding large-scale functional magnetic resonance imaging (fMRI) datasets that provide high-resolution brain responses to thousands of naturalistic visual stimuli. Because such experiments necessarily involve brief stimulus durations and few repetitions of each stimulus, achieving sufficient signal-to-noise ratio can be a major challenge. We address this challenge by introducing GLMsingle, a scalable, user-friendly toolbox available in MATLAB and Python that enables accurate estimation of single-trial fMRI responses (glmsingle.org). Requiring only fMRI time-series data and a design matrix as inputs, GLMsingle integrates three techniques for improving the accuracy of trial-wise general linear model (GLM) beta estimates. First, for each voxel, a custom hemodynamic response function (HRF) is identified from a library of candidate functions. Second, cross-validation is used to derive a set of noise regressors from voxels unrelated to the experiment. Third, to improve the stability of beta estimates for closely spaced trials, betas are regularized on a voxel-wise basis using ridge regression. Applying GLMsingle to the Natural Scenes Dataset and BOLD5000, we find that GLMsingle substantially improves the reliability of beta estimates across visually-responsive cortex in all subjects. Comparable improvements in reliability are also observed in a smaller-scale auditory dataset from the StudyForrest experiment. These improvements translate into tangible benefits for higher-level analyses relevant to systems and cognitive neuroscience. We demonstrate that GLMsingle: (i) helps decorrelate response estimates between trials nearby in time; (ii) enhances representational similarity between subjects within and across datasets; and (iii) boosts one-versus-many decoding of visual stimuli. GLMsingle is a publicly available tool that can significantly improve the quality of past, present, and future neuroimaging datasets sampling brain activity across many experimental conditions.

Mesoscopic in vivo human T2* dataset acquired using quantitative MRI at 7 Tesla.

Mesoscopic (0.1-0.5 mm) interrogation of the living human brain is critical for advancing neuroscience and bridging the resolution gap with animal models. Despite the variety of MRI contrasts measured in recent years at the mesoscopic scale, in vivo quantitative imaging of T2* has not been performed. Here we provide a dataset containing empirical T2* measurements acquired at 0.35 × 0.35 × 0.35 mm3 voxel resolution using 7 Tesla MRI. To demonstrate unique features and high quality of this dataset, we generate flat map visualizations that reveal fine-scale cortical substructures such as layers and vessels, and we report quantitative depth-dependent T2* (as well as R2*) values in primary visual cortex and auditory cortex that are highly consistent across subjects. This dataset is freely available at https://doi.org/10.17605/OSF.IO/N5BJ7, and may prove useful for anatomical investigations of the human brain, as well as for improving our understanding of the basis of the T2*-weighted (f)MRI signal.

Natural scene sampling reveals reliable coarse-scale orientation tuning in human V1.

Orientation selectivity in primate visual cortex is organized into cortical columns. Since cortical columns are at a finer spatial scale than the sampling resolution of standard BOLD fMRI measurements, analysis approaches have been proposed to peer past these spatial resolution limitations. It was recently found that these methods are predominantly sensitive to stimulus vignetting - a form of selectivity arising from an interaction of the oriented stimulus with the aperture edge. Beyond vignetting, it is not clear whether orientation-selective neural responses are detectable in BOLD measurements. Here, we leverage a dataset of visual cortical responses measured using high-field 7T fMRI. Fitting these responses using image-computable models, we compensate for vignetting and nonetheless find reliable tuning for orientation. Results further reveal a coarse-scale map of orientation preference that may constitute the neural basis for known perceptual anisotropies. These findings settle a long-standing debate in human neuroscience, and provide insights into functional organization principles of visual cortex.