A voxel-wise encoding model for early visual areas decodes mental images of remembered scenes

Suffering more in imagination than in reality? Mental imagery and fear generalization

Mental imagery may represent a weaker form of perception and, thus, mental images may be more ambiguous than visual percepts. If correct, the acquisition of fear would be less specific for imagined fears in comparison to perceptual fears, perhaps facilitating broader fear generalization. To test this idea, a two-day differential fear conditioning experiment (N = 98) was conducted. On day one, two groups of participants underwent differential fear conditioning such that a specific Gabor patch orientation (CS+) was paired with mild shocks (US) while a second Gabor patch of orthogonal orientation (CS-) was never paired with shock. Critically, one group imagined the Gabor patches and the other group was visually presented the Gabor patches. Next, both groups were presented visual Gabor patches of similar orientations (GCS) to the CS+. On day two, to assess the persistence of imagined fear, participants returned to the lab and were tested on the GCS devoid of shock. For day one, in contrast to our primary hypothesis, both self-report and skin conductance response measures did not show a significant interaction between the GCS and groups. On day two, both measures demonstrated a persistence of imagined fear, without US delivery. Taken together, rather than demonstrating an overgeneralization effect, the results from this study suggest that imagery-based fear conditioning generalizes to a similar extent as perceptually acquired fear conditioning. Further, the persistence of imagery-based fear may have unique extinction qualities in comparison to perceptual-based fear.

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.

Laminar dissociation of feedforward and feedback in high-level ventral visual cortex during imagery and perception

Visual imagery and perception share neural machinery but rely on different information flow. While perception is driven by the integration of sensory feedforward and internally generated feedback information, imagery relies on feedback only. This suggests that although imagery and perception may activate overlapping brain regions, they do so in informationally distinctive ways. Using lamina-resolved MRI at 7 T, we measured the neural activity during imagery and perception of faces and scenes in high-level ventral visual cortex at the mesoscale of laminar organization that distinguishes feedforward from feedback signals. We found distinctive laminar profiles for imagery and perception of scenes and faces in the parahippocampal place area and the fusiform face area, respectively. Our findings provide insight into the neural basis of the phenomenology of visual imagery versus perception and shed new light into the mesoscale organization of feedforward and feedback information flow in high-level ventral visual cortex.

Uncovering the Role of the Early Visual Cortex in Visual Mental Imagery

The question of whether the early visual cortex (EVC) is involved in visual mental imagery remains a topic of debate. In this paper, I propose that the inconsistency in findings can be explained by the unique challenges associated with investigating EVC activity during imagery. During perception, the EVC processes low-level features, which means that activity is highly sensitive to variation in visual details. If the EVC has the same role during visual mental imagery, any change in the visual details of the mental image would lead to corresponding changes in EVC activity. Within this context, the question should not be whether the EVC is 'active' during imagery but how its activity relates to specific imagery properties. Studies using methods that are sensitive to variation in low-level features reveal that imagery can recruit the EVC in similar ways as perception. However, not all mental images contain a high level of visual details. Therefore, I end by considering a more nuanced view, which states that imagery can recruit the EVC, but that does not mean that it always does so.

Feedback scales the spatial tuning of cortical responses during visual memory

Perception, working memory, and long-term memory each evoke neural responses in visual cortex, suggesting that memory uses encoding mechanisms shared with perception. While previous research has largely focused on how perception and memory are similar, we hypothesized that responses in visual cortex would differ depending on the origins of the inputs. Using fMRI, we quantified spatial tuning in visual cortex while participants (both sexes) viewed, maintained in working memory, or retrieved from long-term memory a peripheral target. In each of these conditions, BOLD responses were spatially tuned and were aligned with the target's polar angle in all measured visual field maps including V1. As expected given the increasing sizes of receptive fields, polar angle tuning during perception increased in width systematically up the visual hierarchy from V1 to V2, V3, hV4, and beyond. In stark contrast, the widths of tuned responses were broad across the visual hierarchy during working memory and long-term memory, matched to the widths in perception in later visual field maps but much broader in V1. This pattern is consistent with the idea that mnemonic responses in V1 stem from top-down sources. Moreover, these tuned responses when biased (clockwise or counterclockwise of target) predicted matched biases in memory, suggesting that the readout of maintained and reinstated mnemonic responses influences memory guided behavior. We conclude that feedback constrains spatial tuning during memory, where earlier visual maps inherit broader tuning from later maps thereby impacting the precision of memory.

Linguistic modulation of the neural encoding of phonemes

Speech comprehension entails the neural mapping of the acoustic speech signal onto learned linguistic units. This acousto-linguistic transformation is bi-directional, whereby higher-level linguistic processes (e.g. semantics) modulate the acoustic analysis of individual linguistic units. Here, we investigated the cortical topography and linguistic modulation of the most fundamental linguistic unit, the phoneme. We presented natural speech and "phoneme quilts" (pseudo-randomly shuffled phonemes) in either a familiar (English) or unfamiliar (Korean) language to native English speakers while recording functional magnetic resonance imaging. This allowed us to dissociate the contribution of acoustic vs. linguistic processes toward phoneme analysis. We show that (i) the acoustic analysis of phonemes is modulated by linguistic analysis and (ii) that for this modulation, both of acoustic and phonetic information need to be incorporated. These results suggest that the linguistic modulation of cortical sensitivity to phoneme classes minimizes prediction error during natural speech perception, thereby aiding speech comprehension in challenging listening situations.

Multisensory subtypes of aphantasia: Mental imagery as supramodal perception in reverse

Cognitive neuroscience research on mental imagery has largely focused on the visual imagery modality in unimodal task contexts. Recent studies have uncovered striking individual differences in visual imagery capacity, with some individuals reporting a subjective absence of conscious visual imagery ability altogether ("aphantasia"). However, naturalistic mental imagery is often multi-sensory, and preliminary findings suggest that many individuals with aphantasia also report a subjective lack of mental imagery in other sensory domains (such as auditory or olfactory imagery). In this paper, we perform a series of cluster analyses on the multi-sensory imagery questionnaire scores of two large groups of aphantasic subjects, defining latent sub-groups in this sample population. We demonstrate that aphantasia is a heterogenous phenomenon characterised by dominant sub-groups of individuals with visual aphantasia (those who report selective visual imagery absence) and multi-sensory aphantasia (those who report an inability to generate conscious mental imagery in any sensory modality). We replicate our findings in a second large sample and show that more unique aphantasia sub-types also exist, such as individuals with selectively preserved mental imagery in only one sensory modality (e.g. intact auditory imagery). We outline the implications of our findings for network theories of mental imagery, discussing how unique aphantasia aetiologies with distinct self-report patterns might reveal alterations to various levels of the sensory processing hierarchy implicated in mental imagery.

Encoded and updated spatial working memories share a common representational format in alpha activity

Working memory (WM) flexibly updates information to adapt to the dynamic environment. Here, we used alpha-band activity in the EEG to reconstruct the content of dynamic WM updates and compared this representational format to static WM content. An inverted encoding model using alpha activity precisely tracked both the initially encoded position and the updated position following an auditory cue signaling mental updating. The timing of the update, as tracked in the EEG, correlated with reaction times and saccade latency. Finally, cross-training analyses revealed a robust generalization of alpha-band reconstruction of WM contents before and after updating. These findings demonstrate that alpha activity tracks the dynamic updates to spatial WM and that the format of this activity is preserved across the encoded and updated representations. Thus, our results highlight a new approach for measuring updates to WM and show common representational formats during dynamic mental updating and static storage.

Working memory signals in early visual cortex are present in weak and strong imagers

It has been suggested that visual images are memorized across brief periods of time by vividly imagining them as if they were still there. In line with this, the contents of both working memory and visual imagery are known to be encoded already in early visual cortex. If these signals in early visual areas were indeed to reflect a combined imagery and memory code, one would predict them to be weaker for individuals with reduced visual imagery vividness. Here, we systematically investigated this question in two groups of participants. Strong and weak imagers were asked to remember images across brief delay periods. We were able to reliably reconstruct the memorized stimuli from early visual cortex during the delay. Importantly, in contrast to the prediction, the quality of reconstruction was equally accurate for both strong and weak imagers. The decodable information also closely reflected behavioral precision in both groups, suggesting it could contribute to behavioral performance, even in the extreme case of completely aphantasic individuals. Our data thus suggest that working memory signals in early visual cortex can be present even in the (near) absence of phenomenal imagery.

Mental image reconstruction from human brain activity: Neural decoding of mental imagery via deep neural network-based Bayesian estimation

Visual images observed by humans can be reconstructed from their brain activity. However, the visualization (externalization) of mental imagery is challenging. Only a few studies have reported successful visualization of mental imagery, and their visualizable images have been limited to specific domains such as human faces or alphabetical letters. Therefore, visualizing mental imagery for arbitrary natural images stands as a significant milestone. In this study, we achieved this by enhancing a previous method. Specifically, we demonstrated that the visual image reconstruction method proposed in the seminal study by Shen et al. (2019) heavily relied on low-level visual information decoded from the brain and could not efficiently utilize the semantic information that would be recruited during mental imagery. To address this limitation, we extended the previous method to a Bayesian estimation framework and introduced the assistance of semantic information into it. Our proposed framework successfully reconstructed both seen images (i.e., those observed by the human eye) and imagined images from brain activity. Quantitative evaluation showed that our framework could identify seen and imagined images highly accurately compared to the chance accuracy (seen: 90.7%, imagery: 75.6%, chance accuracy: 50.0%). In contrast, the previous method could only identify seen images (seen: 64.3%, imagery: 50.4%). These results suggest that our framework would provide a unique tool for directly investigating the subjective contents of the brain such as illusions, hallucinations, and dreams.

How does the primate brain combine generative and discriminative computations in vision?

Vision is widely understood as an inference problem. However, two contrasting conceptions of the inference process have each been influential in research on biological vision as well as the engineering of machine vision. The first emphasizes bottom-up signal flow, describing vision as a largely feedforward, discriminative inference process that filters and transforms the visual information to remove irrelevant variation and represent behaviorally relevant information in a format suitable for downstream functions of cognition and behavioral control. In this conception, vision is driven by the sensory data, and perception is direct because the processing proceeds from the data to the latent variables of interest. The notion of "inference" in this conception is that of the engineering literature on neural networks, where feedforward convolutional neural networks processing images are said to perform inference. The alternative conception is that of vision as an inference process in Helmholtz's sense, where the sensory evidence is evaluated in the context of a generative model of the causal processes that give rise to it. In this conception, vision inverts a generative model through an interrogation of the sensory evidence in a process often thought to involve top-down predictions of sensory data to evaluate the likelihood of alternative hypotheses. The authors include scientists rooted in roughly equal numbers in each of the conceptions and motivated to overcome what might be a false dichotomy between them and engage the other perspective in the realm of theory and experiment. The primate brain employs an unknown algorithm that may combine the advantages of both conceptions. We explain and clarify the terminology, review the key empirical evidence, and propose an empirical research program that transcends the dichotomy and sets the stage for revealing the mysterious hybrid algorithm of primate vision.

Prefrontal Cortex Codes Representations of Target Identity and Feature Uncertainty

Many objects in the real world have features that vary over time, creating uncertainty in how they will look in the future. This uncertainty makes statistical knowledge about the likelihood of features critical to attention demanding processes such as visual search. However, little is known about how the uncertainty of visual features is integrated into predictions about search targets in the brain. In the current study, we test the idea that regions prefrontal cortex code statistical knowledge about search targets before the onset of search. Across 20 human participants (13 female; 7 male), we observe target identity in the multivariate pattern and uncertainty in the overall activation of dorsolateral prefrontal cortex (DLPFC) and inferior frontal junction (IFJ) in advance of the search display. This indicates that the target identity (mean) and uncertainty (variance) of the target distribution are coded independently within the same regions. Furthermore, once the search display appears the univariate IFJ signal scaled with the distance of the actual target from the expected mean, but more so when expected variability was low. These results inform neural theories of attention by showing how the prefrontal cortex represents both the identity and expected variability of features in service of top-down attentional control.SIGNIFICANCE STATEMENT Theories of attention and working memory posit that when we engage in complex cognitive tasks our performance is determined by how precisely we remember task-relevant information. However, in the real world the properties of objects change over time, creating uncertainty about many aspects of the task. There is currently a gap in our understanding of how neural systems represent this uncertainty and combine it with target identity information in anticipation of attention demanding cognitive tasks. In this study, we show that the prefrontal cortex represents identity and uncertainty as unique codes before task onset. These results advance theories of attention by showing that the prefrontal cortex codes both target identity and uncertainty to implement top-down attentional control.

Mental imagery and visual attentional templates: A dissociation

There is a growing interest in the relationship between mental images and attentional templates as both are considered pictorial representations that involve similar neural mechanisms. Here, we investigated the role of mental imagery in the automatic implementation of attentional templates and their effect on involuntary attention. We developed a novel version of the contingent capture paradigm designed to encourage the generation of a new template on each trial and measure contingent spatial capture by a template-matching visual feature (color). Participants were required to search at four different locations for a specific object indicated at the start of each trial. Immediately prior to the search display, color cues were presented surrounding the potential target locations, one of which matched the target color (e.g., red for strawberry). Across three experiments, our task induced a robust contingent capture effect, reflected by faster responses when the target appeared in the location previously occupied by the target-matching cue. Contrary to our predictions, this effect remained consistent regardless of self-reported individual differences in visual mental imagery (Experiment 1, N = 216) or trial-by-trial variation of voluntary imagery vividness (Experiment 2, N = 121). Moreover, contingent capture was observed even among aphantasic participants, who report no imagery (Experiment 3, N = 91). The magnitude of the effect was not reduced in aphantasics compared to a control sample of non-aphantasics, although the two groups reported substantial differences in their search strategy and exhibited differences in overall speed and accuracy. Our results hence establish a dissociation between the generation and implementation of attentional templates for a visual feature (color) and subjectively experienced imagery.

Different Mechanisms for Supporting Mental Imagery and Perceptual Representations: Modulation Versus Excitation

Recent research suggests imagery is functionally equivalent to a weak form of visual perception. Here we report evidence across five independent experiments on adults that perception and imagery are supported by fundamentally different mechanisms: Whereas perceptual representations are largely formed via increases in excitatory activity, imagery representations are largely supported by modulating nonimagined content. We developed two behavioral techniques that allowed us to first put the visual system into a state of adaptation and then probe the additivity of perception and imagery. If imagery drives similar excitatory visual activity to perception, pairing imagery with perceptual adapters should increase the state of adaptation. Whereas pairing weak perception with adapters increased measures of adaptation, pairing imagery reversed their effects. Further experiments demonstrated that these nonadditive effects were due to imagery weakening representations of nonimagined content. Together these data provide empirical evidence that the brain uses categorically different mechanisms to represent imagery and perception.

Visualizing Compassion: Episodic Simulation as Contemplative Practice

Contemplative interventions designed to cultivate compassion are receiving increasing empirical attention. Accumulating evidence suggests that these interventions bolster prosocial motivation and warmth towards others. Less is known about how these practices impact compassion in everyday life. Here we consider one mechanistic pathway through which compassion practices may impact perception and action in the world: simulation. Evidence suggests that vividly imagining a situation simulates that experience in the brain as if it were, to a degree, actually happening. Thus, we hypothesize that simulation during imagery-based contemplative practices can construct sensorimotor patterns in the brain that prime an individual to act compassionately in the world. We first present evidence across multiple literatures in Psychology that motivates this hypothesis, including the neuroscience of mental imagery and the emerging literature on prosocial episodic simulation. Then, we examine the specific contemplative practices in compassion-based interventions that may construct such simulations. We conclude with future directions for investigating how compassion-based interventions may shape prosocial perception and action in everyday life.

Adding Meaning to Memories: How Parietal Cortex Combines Semantic Content with Episodic Experience

Neuroimaging studies of human memory have consistently found that univariate responses in parietal cortex track episodic experience with stimuli (whether stimuli are 'old' or 'new'). More recently, pattern-based fMRI studies have shown that parietal cortex also carries information about the semantic content of remembered experiences. However, it is not well understood how memory-based and content-based signals are integrated within parietal cortex. Here, in humans (males and females), we used voxel-wise encoding models and a recognition memory task to predict the fMRI activity patterns evoked by complex natural scene images based on (1) the episodic history and (2) the semantic content of each image. Models were generated and compared across distinct subregions of parietal cortex and for occipitotemporal cortex. We show that parietal and occipitotemporal regions each encode memory and content information, but they differ in how they combine this information. Among parietal subregions, angular gyrus was characterized by robust and overlapping effects of memory and content. Moreover, subject-specific semantic tuning functions revealed that successful recognition shifted the amplitude of tuning functions in angular gyrus but did not change the selectivity of tuning. In other words, effects of memory and content were additive in angular gyrus. This pattern of data contrasted with occipitotemporal cortex where memory and content effects were interactive: memory effects were preferentially expressed by voxels tuned to the content of a remembered image. Collectively, these findings provide unique insight into how parietal cortex combines information about episodic memory and semantic content.SIGNIFICANCE STATEMENT Neuroimaging studies of human memory have identified multiple brain regions that not only carry information about "whether" a visual stimulus is successfully recognized but also "what" the content of that stimulus includes. However, a fundamental and open question concerns how the brain integrates these two types of information (memory and content). Here, using a powerful combination of fMRI analysis methods, we show that parietal cortex, particularly the angular gyrus, robustly combines memory- and content-related information, but these two forms of information are represented via additive, independent signals. In contrast, memory effects in high-level visual cortex critically depend on (and interact with) content representations. Together, these findings reveal multiple and distinct ways in which the brain combines memory- and content-related information.

Visual Representations: Insights from Neural Decoding

Patterns of brain activity contain meaningful information about the perceived world. Recent decades have welcomed a new era in neural analyses, with computational techniques from machine learning applied to neural data to decode information represented in the brain. In this article, we review how decoding approaches have advanced our understanding of visual representations and discuss efforts to characterize both the complexity and the behavioral relevance of these representations. We outline the current consensus regarding the spatiotemporal structure of visual representations and review recent findings that suggest that visual representations are at once robust to perturbations, yet sensitive to different mental states. Beyond representations of the physical world, recent decoding work has shone a light on how the brain instantiates internally generated states, for example, during imagery and prediction. Going forward, decoding has remarkable potential to assess the functional relevance of visual representations for human behavior, reveal how representations change across development and during aging, and uncover their presentation in various mental disorders.

Mapping multidimensional content representations to neural and behavioral expressions of episodic memory

Human neuroimaging studies have shown that the contents of episodic memories are represented in distributed patterns of neural activity. However, these studies have mostly been limited to decoding simple, unidimensional properties of stimuli. Semantic encoding models, in contrast, offer a means for characterizing the rich, multidimensional information that comprises episodic memories. Here, we extensively sampled four human fMRI subjects to build semantic encoding models and then applied these models to reconstruct content from natural scene images as they were viewed and recalled from memory. First, we found that multidimensional semantic information was successfully reconstructed from activity patterns across visual and lateral parietal cortices, both when viewing scenes and when recalling them from memory. Second, whereas visual cortical reconstructions were much more accurate when images were viewed versus recalled from memory, lateral parietal reconstructions were comparably accurate across visual perception and memory. Third, by applying natural language processing methods to verbal recall data, we showed that fMRI-based reconstructions reliably matched subjects' verbal descriptions of their memories. In fact, reconstructions from ventral temporal cortex more closely matched subjects' own verbal recall than other subjects' verbal recall of the same images. Fourth, encoding models reliably transferred across subjects: memories were successfully reconstructed using encoding models trained on data from entirely independent subjects. Together, these findings provide evidence for successful reconstructions of multidimensional and idiosyncratic memory representations and highlight the differential sensitivity of visual cortical and lateral parietal regions to information derived from the external visual environment versus internally-generated memories.

MIRACLE: MInd ReAding CLassification Engine

Brain-computer interfaces (BCIs) have revolutionized the way humans interact with machines, particularly for patients with severe motor impairments. EEG-based BCIs have limited functionality due to the restricted pool of stimuli that they can distinguish, while those elaborating event-related potentials up to now employ paradigms that require the patient's perception of the eliciting stimulus. In this work, we propose MIRACLE: a novel BCI system that combines functional data analysis and machine-learning techniques to decode patients' minds from the elicited potentials. MIRACLE relies on a hierarchical ensemble classifier recognizing 10 different semantic categories of imagined stimuli. We validated MIRACLE on an extensive dataset collected from 20 volunteers, with both imagined and perceived stimuli, to compare the system performance on the two. Furthermore, we quantify the importance of each EEG channel in the decision-making process of the classifier, which can help reduce the number of electrodes required for data acquisition, enhancing patients' comfort.

A comparison of mental imagery and perceptual cueing across domains of attention

Mental imagery and perceptual cues can influence subsequent visual search performance, but examination of this influence has been limited to low-level features like colors and shapes. The present study investigated how the two types of cues influence low-level visual search, visual search with realistic objects, and executive attention. On each trial, participants were either presented with a colored square or tasked with using mental imagery to generate a colored square that could match the target (valid trial) or distractor (invalid trial) in the search array that followed (Experiments 1 and 3). In a separate experiment, the colored square displayed or generated was replaced with a realistic object in a specific category that could appear as a target or distractor in the search array (Experiment 2). Although the displayed object was in the same category as an item in the search display, they were never a perfect match (e.g., jam drop cookie instead of chocolate chip). Our findings revealed that the facilitation of performance on valid trials compared with invalid trials was greater for perceptual cues than imagery cues for low-level features (Experiment 1), whereas the influence of these two types of cues was comparable in the context of realistic objects (Experiment 2) The influence of mental imagery appears not to extend to the resolution of conflict generated by color-word Stroop stimuli (Experiment 3). The present findings extend our understanding of how mental imagery influences the allocation of attention.

Vividness as the similarity between generated imagery and an internal model

Vividness in visual mental imagery has been relatively under-explored compared to imagery's representational format and neural mechanisms. In this paper, we take a deeper look at vividness and suggest that in re-framing it, we can potentially reconcile disparate findings regarding visual cortex activation during imagery. Unlike traditional views of vividness that define the concept in terms of perception, we frame vividness in terms of imagery's relation to an internal model; the closer the generated imagery is to this model, the more vivid it is. This view is considered alongside existing neuroscientific, psychological, and philosophical research, as well as directions for future research.

Semantic reconstruction of continuous language from non-invasive brain recordings

A brain-computer interface that decodes continuous language from non-invasive recordings would have many scientific and practical applications. Currently, however, non-invasive language decoders can only identify stimuli from among a small set of words or phrases. Here we introduce a non-invasive decoder that reconstructs continuous language from cortical semantic representations recorded using functional magnetic resonance imaging (fMRI). Given novel brain recordings, this decoder generates intelligible word sequences that recover the meaning of perceived speech, imagined speech and even silent videos, demonstrating that a single decoder can be applied to a range of tasks. We tested the decoder across cortex and found that continuous language can be separately decoded from multiple regions. As brain-computer interfaces should respect mental privacy, we tested whether successful decoding requires subject cooperation and found that subject cooperation is required both to train and to apply the decoder. Our findings demonstrate the viability of non-invasive language brain-computer interfaces.

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.

What do you have in mind? ERP markers of visual and auditory imagery

This study aimed to investigate the psychophysiological markers of imagery processes through EEG/ERP recordings. Visual and auditory stimuli representing 10 different semantic categories were shown to 30 healthy participants. After a given interval and prompted by a light signal, participants were asked to activate a mental image corresponding to the semantic category for recording synchronized electrical potentials. Unprecedented electrophysiological markers of imagination were recorded in the absence of sensory stimulation. The following peaks were identified at specific scalp sites and latencies, during imagination of infants (centroparietal positivity, CPP, and late CPP), human faces (anterior negativity, AN), animals (anterior positivity, AP), music (P300-like), speech (N400-like), affective vocalizations (P2-like) and sensory (visual vs auditory) modality (PN300). Overall, perception and imagery conditions shared some common electro/cortical markers, but during imagery the category-dependent modulation of ERPs was long latency and more anterior, with respect to the perceptual condition. These ERP markers might be precious tools for BCI systems (pattern recognition, classification, or A.I. algorithms) applied to patients affected by consciousness disorders (e.g., in a vegetative or comatose state) or locked-in-patients (e.g., spinal or SLA patients).

Seeing the future: Connectome strength and network efficiency in visual network predict individual ability of episodic future thinking

Episodic future thinking (EFT) refers to the critical ability that enables people to construct and pre-experience the vivid mental imagery about future events, which impacts on the decision-making for individuals and group. Although EFT is generally believed to have a visual nature by theorists, little neuroscience evidence has been provided to verify this assumption. Here, by employing the approach of connectome-based predictive modeling (CPM) and graph-theoretical analysis, we analyzed resting-state functional brain image from 191 participants to predict their variability of EFT ability (leave-one-out cross-validation), and validated the results by applying different parcellation schemas and feature selection thresholds. At the connectome strength level, CPM-based analysis revealed that EFT ability could be predicted by the connectome strength of visual network. Besides, at the network level, graph-theoretical analysis showed that EFT ability could be predicted by the network efficiency of visual network. Moreover, these findings were replicated using different parcellation schemas and feature selection thresholds. These results robustly and collectively supported that the visual network might be one of the neural substrates underlying EFT ability from a comprehensive perspective of resting-state functional connectivity strength and the neural network. This study provides indications on how the function of visual network supports EFT ability, and enhances our understanding of the EFT ability from a neural basis perspective.

A generative adversarial model of intrusive imagery in the human brain

The mechanisms underlying the subjective experiences of mental disorders remain poorly understood. This is partly due to long-standing over-emphasis on behavioral and physiological symptoms and a de-emphasis of the patient's subjective experiences when searching for treatments. Here, we provide a new perspective on the subjective experience of mental disorders based on findings in neuroscience and artificial intelligence (AI). Specifically, we propose the subjective experience that occurs in visual imagination depends on mechanisms similar to generative adversarial networks that have recently been developed in AI. The basic idea is that a generator network fabricates a prediction of the world, and a discriminator network determines whether it is likely real or not. Given that similar adversarial interactions occur in the two major visual pathways of perception in people, we explored whether we could leverage this AI-inspired approach to better understand the intrusive imagery experiences of patients suffering from mental illnesses such as post-traumatic stress disorder (PTSD) and acute stress disorder. In our model, a nonconscious visual pathway generates predictions of the environment that influence the parallel but interacting conscious pathway. We propose that in some patients, an imbalance in these adversarial interactions leads to an overrepresentation of disturbing content relative to current reality, and results in debilitating flashbacks. By situating the subjective experience of intrusive visual imagery in the adversarial interaction of these visual pathways, we propose testable hypotheses on novel mechanisms and clinical applications for controlling and possibly preventing symptoms resulting from intrusive imagery.

A large and rich EEG dataset for modeling human visual object recognition

The human brain achieves visual object recognition through multiple stages of linear and nonlinear transformations operating at a millisecond scale. To predict and explain these rapid transformations, computational neuroscientists employ machine learning modeling techniques. However, state-of-the-art models require massive amounts of data to properly train, and to the present day there is a lack of vast brain datasets which extensively sample the temporal dynamics of visual object recognition. Here we collected a large and rich dataset of high temporal resolution EEG responses to images of objects on a natural background. This dataset includes 10 participants, each with 82,160 trials spanning 16,740 image conditions. Through computational modeling we established the quality of this dataset in five ways. First, we trained linearizing encoding models that successfully synthesized the EEG responses to arbitrary images. Second, we correctly identified the recorded EEG data image conditions in a zero-shot fashion, using EEG synthesized responses to hundreds of thousands of candidate image conditions. Third, we show that both the high number of conditions as well as the trial repetitions of the EEG dataset contribute to the trained models' prediction accuracy. Fourth, we built encoding models whose predictions well generalize to novel participants. Fifth, we demonstrate full end-to-end training of randomly initialized DNNs that output EEG responses for arbitrary input images. We release this dataset as a tool to foster research in visual neuroscience and computer vision.

Representations and decodability of diverse cognitive functions are preserved across the human cortex, cerebellum, and subcortex

Which part of the brain contributes to our complex cognitive processes? Studies have revealed contributions of the cerebellum and subcortex to higher-order cognitive functions; however, it has been unclear whether such functional representations are preserved across the cortex, cerebellum, and subcortex. In this study, we use functional magnetic resonance imaging data with 103 cognitive tasks and construct three voxel-wise encoding and decoding models independently using cortical, cerebellar, and subcortical voxels. Representational similarity analysis reveals that the structure of task representations is preserved across the three brain parts. Principal component analysis visualizes distinct organizations of abstract cognitive functions in each part of the cerebellum and subcortex. More than 90% of the cognitive tasks are decodable from the cerebellum and subcortical activities, even for the novel tasks not included in model training. Furthermore, we show that the cerebellum and subcortex have sufficient information to reconstruct activity in the cerebral cortex.

Experimental evidence for involvement of monocular channels in mental rotation

According to the prevailing view, cognitive processes of mental rotation are carried out by visuospatial perceptual circuits located primarily in high cortical areas. Here, we examined the functional involvement of (mostly subcortical) monocular channels in mental rotation tasks. Images of two rotated objects (0°, 50°, 100°, or 150°; identical or mirrored) were presented either to one eye (monocular) or segregated between the eyes (interocular). The results indicated a causal role for low monocular visual channels in mental rotation: Response times for identical ("same") objects at high angular disparities (100°, 150°) were shorter when both objects were presented to a single eye than when each object was presented to a different eye. We suggest that mental rotation processes rely on cortico-subcortical loops that support visuospatial perception. More generally, the findings highlight the potential contribution of lower-level mechanisms to what are typically considered to be high-level cognitive functions, such as mental representation.

Perception and memory have distinct spatial tuning properties in human visual cortex

Reactivation of earlier perceptual activity is thought to underlie long-term memory recall. Despite evidence for this view, it is unclear whether mnemonic activity exhibits the same tuning properties as feedforward perceptual activity. Here, we leverage population receptive field models to parameterize fMRI activity in human visual cortex during spatial memory retrieval. Though retinotopic organization is present during both perception and memory, large systematic differences in tuning are also evident. Whereas there is a three-fold decline in spatial precision from early to late visual areas during perception, this pattern is not observed during memory retrieval. This difference cannot be explained by reduced signal-to-noise or poor performance on memory trials. Instead, by simulating top-down activity in a network model of cortex, we demonstrate that this property is well explained by the hierarchical structure of the visual system. Together, modeling and empirical results suggest that computational constraints imposed by visual system architecture limit the fidelity of memory reactivation in sensory cortex.

Neural activity from perception is thought to be reactivated during memory recall. Here, the authors show that memory reactivation in visual cortex exhibits different tuning properties from those observed during perception.

Decoding of Motor Imagery Involving Whole-body Coordination

The present study investigated whether different types of motor imageries can be classified based on the location of the activation peaks or the multivariate pattern analysis (MVPA) of functional magnetic resonance imaging (fMRI) and compared the difference between visual motor imagery (VI) and kinesthetic motor imagery (KI). During fMRI scanning sessions, 25 participants imagined four movements included in the Motor Imagery Questionnaire-Revised (MIQ-R): knee lift, jump, arm movement, and waist bend. These four imagined movements were then classified based on the peak location or the patterns of fMRI signal values. We divided the participants into two groups based on whether they found it easier to generate VI (VI group, n = 10) or KI (KI group, n = 15). Our results show that the imagined movements can be classified using both the location of the activation peak and the spatial activation patterns within the sensorimotor cortex, and MVPA performs better than the activation peak classification. Furthermore, our result reveals that the KI group achieved a higher MVPA decoding accuracy within the left primary somatosensory cortex than the VI group, suggesting that the modality of motor imagery differently affects the classification performance in distinct brain regions.

Self-supervised Natural Image Reconstruction and Large-scale Semantic Classification from Brain Activity

Reconstructing natural images and decoding their semantic category from fMRI brain recordings is challenging. Acquiring sufficient pairs of images and their corresponding fMRI responses, which span the huge space of natural images, is prohibitive. We present a novel self-supervised approach that goes well beyond the scarce paired data, for achieving both: (i) state-of-the art fMRI-to-image reconstruction, and (ii) first-ever large-scale semantic classification from fMRI responses. By imposing cycle consistency between a pair of deep neural networks (from image-to-fMRI & from fMRI-to-image), we train our image reconstruction network on a large number of "unpaired" natural images (images without fMRI recordings) from many novel semantic categories. This enables to adapt our reconstruction network to a very rich semantic coverage without requiring any explicit semantic supervision. Specifically, we find that combining our self-supervised training with high-level perceptual losses, gives rise to new reconstruction & classification capabilities. In particular, this perceptual training enables to classify well fMRIs of never-before-seen semantic classes, without requiring any class labels during training. This gives rise to: (i) Unprecedented image-reconstruction from fMRI of never-before-seen images (evaluated by image metrics and human testing), and (ii) Large-scale semantic classification of categories that were never-before-seen during network training. Such large-scale (1000-way) semantic classification from fMRI recordings has never been demonstrated before. Finally, we provide evidence for the biological consistency of our learned model.

Confidence modulates the decodability of scene prediction during partially-observable maze exploration in humans

Prediction ability often involves some degree of uncertainty-a key determinant of confidence. Here, we sought to assess whether predictions are decodable in partially-observable environments where one's state is uncertain, and whether this information is sensitive to confidence produced by such uncertainty. We used functional magnetic resonance imaging-based, partially-observable maze navigation tasks in which subjects predicted upcoming scenes and reported their confidence regarding these predictions. Using a multi-voxel pattern analysis, we successfully decoded both scene predictions and subjective confidence from activities in the localized parietal and prefrontal regions. We also assessed confidence in their beliefs about where they were in the maze. Importantly, prediction decodability varied according to subjective scene confidence in the superior parietal lobule and state confidence estimated by the behavioral model in the inferior parietal lobule. These results demonstrate that prediction in uncertain environments depends on the prefrontal-parietal network within which prediction and confidence interact.

The pupillary light response as a physiological index of aphantasia, sensory and phenomenological imagery strength

The pupillary light response is an important automatic physiological response which optimizes light reaching the retina. Recent work has shown that the pupil also adjusts in response to illusory brightness and a range of cognitive functions, however, it remains unclear what exactly drives these endogenous changes. Here, we show that the imagery pupillary light response correlates with objective measures of sensory imagery strength. Further, the trial-by-trial phenomenological vividness of visual imagery is tracked by the imagery pupillary light response. We also demonstrated that a group of individuals without visual imagery (aphantasia) do not show any significant evidence of an imagery pupillary light response, however they do show perceptual pupil light responses and pupil dilation with larger cognitive load. Our results provide evidence that the pupillary light response indexes the sensory strength of visual imagery. This work also provides the first physiological validation of aphantasia.

Indicators and criteria of consciousness: ethical implications for the care of behaviourally unresponsive patients

Background

Assessing consciousness in other subjects, particularly in non-verbal and behaviourally disabled subjects (e.g., patients with disorders of consciousness), is notoriously challenging but increasingly urgent. The high rate of misdiagnosis among disorders of consciousness raises the need for new perspectives in order to inspire new technical and clinical approaches.

Main body

We take as a starting point a recently introduced list of operational indicators of consciousness that facilitates its recognition in challenging cases like non-human animals and Artificial Intelligence to explore their relevance to disorders of consciousness and their potential ethical impact on the diagnosis and healthcare of relevant patients. Indicators of consciousness mean particular capacities that can be deduced from observing the behaviour or cognitive performance of the subject in question (or from neural correlates of such performance) and that do not define a hard threshold in deciding about the presence of consciousness, but can be used to infer a graded measure based on the consistency amongst the different indicators. The indicators of consciousness under consideration offer a potential useful strategy for identifying and assessing residual consciousness in patients with disorders of consciousness, setting the theoretical stage for an operationalization and quantification of relevant brain activity.

Conclusions

Our heuristic analysis supports the conclusion that the application of the identified indicators of consciousness to its disorders will likely inspire new strategies for assessing three very urgent issues: the misdiagnosis of disorders of consciousness; the need for a gold standard in detecting consciousness and diagnosing its disorders; and the need for a refined taxonomy of disorders of consciousness.

Spatiotemporally resolved multivariate pattern analysis for M/EEG

An emerging goal in neuroscience is tracking what information is represented in brain activity over time as a participant completes some task. While electroencephalography (EEG) and magnetoencephalography (MEG) offer millisecond temporal resolution of how activity patterns emerge and evolve, standard decoding methods present significant barriers to interpretability as they obscure the underlying spatial and temporal activity patterns. We instead propose the use of a generative encoding model framework that simultaneously infers the multivariate spatial patterns of activity and the variable timing at which these patterns emerge on individual trials. An encoding model inversion maps from these parameters to the equivalent decoding model, allowing predictions to be made about unseen test data in the same way as in standard decoding methodology. These SpatioTemporally Resolved MVPA (STRM) models can be flexibly applied to a wide variety of experimental paradigms, including classification and regression tasks. We show that these models provide insightful maps of the activity driving predictive accuracy metrics; demonstrate behaviourally meaningful variation in the timing of pattern emergence on individual trials; and achieve predictive accuracies that are either equivalent or surpass those achieved by more widely used methods. This provides a new avenue for investigating the brain's representational dynamics and could ultimately support more flexible experimental designs in the future.

Voluntary control of semantic neural representations by imagery with conflicting visual stimulation

Neural representations of visual perception are affected by mental imagery and attention. Although attention is known to modulate neural representations, it is unknown how imagery changes neural representations when imagined and perceived images semantically conflict. We hypothesized that imagining an image would activate a neural representation during its perception even while watching a conflicting image. To test this hypothesis, we developed a closed-loop system to show images inferred from electrocorticograms using a visual semantic space. The successful control of the feedback images demonstrated that the semantic vector inferred from electrocorticograms became closer to the vector of the imagined category, even while watching images from different categories. Moreover, modulation of the inferred vectors by mental imagery depended asymmetrically on the perceived and imagined categories. Shared neural representation between mental imagery and perception was still activated by the imagery under semantically conflicting perceptions depending on the semantic category.

In this study, intracranial EEG recordings show that neural representations of imagined images can still be present in humans even when they are shown conflicting images.

Perceptual reality monitoring: Neural mechanisms dissociating imagination from reality

There is increasing evidence that imagination relies on similar neural mechanisms as externally triggered perception. This overlap presents a challenge for perceptual reality monitoring: deciding what is real and what is imagined. Here, we explore how perceptual reality monitoring might be implemented in the brain. We first describe sensory and cognitive factors that could dissociate imagery and perception and conclude that no single factor unambiguously signals whether an experience is internally or externally generated. We suggest that reality monitoring is implemented by higher-level cortical circuits that evaluate first-order sensory and cognitive factors to determine the source of sensory signals. According to this interpretation, perceptual reality monitoring shares core computations with metacognition. This multi-level architecture might explain several types of source confusion as well as dissociations between simply knowing whether something is real and actually experiencing it as real. We discuss avenues for future research to further our understanding of perceptual reality monitoring, an endeavour that has important implications for our understanding of clinical symptoms as well as general cognitive function.

Relative precision of top-down attentional modulations is lower in early visual cortex compared to mid- and high-level visual areas

Top-down spatial attention enhances cortical representations of behaviorally relevant visual information and increases the precision of perceptual reports. However, little is known about the relative precision of top-down attentional modulations in different visual areas, especially compared with the highly precise stimulus-driven responses that are observed in early visual cortex. For example, the precision of attentional modulations in early visual areas may be limited by the relatively coarse spatial selectivity and the anatomical connectivity of the areas in prefrontal cortex that generate and relay the top-down signals. Here, we used functional MRI (fMRI) and human participants to assess the precision of bottom-up spatial representations evoked by high-contrast stimuli across the visual hierarchy. Then, we examined the relative precision of top-down attentional modulations in the absence of spatially specific bottom-up drive. Whereas V1 showed the largest relative difference between the precision of top-down attentional modulations and the precision of bottom-up modulations, midlevel areas such as V4 showed relatively smaller differences between the precision of top-down and bottom-up modulations. Overall, this interaction between visual areas (e.g., V1 vs. V4) and the relative precision of top-down and bottom-up modulations suggests that the precision of top-down attentional modulations is limited by the representational fidelity of areas that generate and relay top-down feedback signals.NEW & NOTEWORTHY When the relative precision of purely top-down and bottom-up signals were compared across visual areas, early visual areas like V1 showed higher bottom-up precision compared with top-down precision. In contrast, midlevel areas showed similar levels of top-down and bottom-up precision. This result suggests that the precision of top-down attentional modulations may be limited by the relatively coarse spatial selectivity and the anatomical connectivity of the areas generating and relaying the signals.

Reconstructing Rapid Natural Vision with fMRI-Conditional Video Generative Adversarial Network

Recent functional magnetic resonance imaging (fMRI) studies have made significant progress in reconstructing perceived visual content, which advanced our understanding of the visual mechanism. However, reconstructing dynamic natural vision remains a challenge because of the limitation of the temporal resolution of fMRI. Here, we developed a novel fMRI-conditional video generative adversarial network (f-CVGAN) to reconstruct rapid video stimuli from evoked fMRI responses. In this model, we employed a generator to produce spatiotemporal reconstructions and employed two separate discriminators (spatial and temporal discriminators) for the assessment. We trained and tested the f-CVGAN on two publicly available video-fMRI datasets, and the model produced pixel-level reconstructions of 8 perceived video frames from each fMRI volume. Experimental results showed that the reconstructed videos were fMRI-related and captured important spatial and temporal information of the original stimuli. Moreover, we visualized the cortical importance map and found that the visual cortex is extensively involved in the reconstruction, whereas the low-level visual areas (V1/V2/V3/V4) showed the largest contribution. Our work suggests that slow blood oxygen level-dependent signals describe neural representations of the fast perceptual process that can be decoded in practice.

Distillation of Regional Activity Reveals Hidden Content of Neural Information in Visual Processing

Multivoxel pattern analysis (MVPA) has become a standard tool for decoding mental states from brain activity patterns. Recent studies have demonstrated that MVPA can be applied to decode activity patterns of a certain region from those of the other regions. By applying a similar region-to-region decoding technique, we examined whether the information represented in the visual areas can be explained by those represented in the other visual areas. We first predicted the brain activity patterns of an area on the visual pathway from the others, then subtracted the predicted patterns from their originals. Subsequently, the visual features were derived from these residuals. During the visual perception task, the elimination of the top-down signals enhanced the simple visual features represented in the early visual cortices. By contrast, the elimination of the bottom-up signals enhanced the complex visual features represented in the higher visual cortices. The directions of such modulation effects varied across visual perception/imagery tasks, indicating that the information flow across the visual cortices is dynamically altered, reflecting the contents of visual processing. These results demonstrated that the distillation approach is a useful tool to estimate the hidden content of information conveyed across brain regions.

Visuospatial coding as ubiquitous scaffolding for human cognition

For more than 100 years we have known that the visual field is mapped onto the surface of visual cortex, imposing an inherently spatial reference frame on visual information processing. Recent studies highlight visuospatial coding not only throughout visual cortex, but also brain areas not typically considered visual. Such widespread access to visuospatial coding raises important questions about its role in wider cognitive functioning. Here, we synthesise these recent developments and propose that visuospatial coding scaffolds human cognition by providing a reference frame through which neural computations interface with environmental statistics and task demands via perception-action loops.

Visual working memory in aphantasia: Retained accuracy and capacity with a different strategy

Visual working memory paradigms involve retaining and manipulating visual information in mind over a period of seconds. Evidence suggests that visual imagery (sensory recruitment) is a strategy used by many to retain visual information during such tasks, leading some researchers to propose that visual imagery and visual working memory may be one and the same. If visual imagery is essential to visual working memory task performance there should be large ramifications for a special population of individuals who do not experience visual imagery, aphantasia. Here we assessed visual working memory task performance in this population using a number of different lab and clinical working memory tasks. We found no differences in capacity limits for visual, general number or spatial working memory for aphantasic individuals compared to controls. Further, aphantasic individuals showed no significant differences in performance on visual components of clinical working memory tests as compared to verbal components. However, there were significant differences in the reported strategies used by aphantasic individuals across all memory tasks. Additionally, aphantasic individual's visual memory accuracy did not demonstrate a significant oblique orientation effect, which is proposed to occur due to sensory recruitment, further supporting their non-visual imagery strategy reports. Taken together these data demonstrate that aphantasic individuals are not impaired on visual working memory tasks, suggesting visual imagery and working memory are not one and the same, with imagery (and sensory recruitment) being just one of the tools that can be used to solve visual working memory tasks.

A neural decoding algorithm that generates language from visual activity evoked by natural images

Transforming neural activities into language is revolutionary for human-computer interaction as well as functional restoration of aphasia. Present rapid development of artificial intelligence makes it feasible to decode the neural signals of human visual activities. In this paper, a novel Progressive Transfer Language Decoding Model (PT-LDM) is proposed to decode visual fMRI signals into phrases or sentences when natural images are being watched. The PT-LDM consists of an image-encoder, a fMRI encoder and a language-decoder. The results showed that phrases and sentences were successfully generated from visual activities. Similarity analysis showed that three often-used evaluation indexes BLEU, ROUGE and CIDEr reached 0.182, 0.197 and 0.680 averagely between the generated texts and the corresponding annotated texts in the testing set respectively, significantly higher than the baseline. Moreover, we found that higher visual areas usually had better performance than lower visual areas and the contribution curve of visual response patterns in language decoding varied at successively different time points. Our findings demonstrate that the neural representations elicited in visual cortices when scenes are being viewed have already contained semantic information that can be utilized to generate human language. Our study shows potential application of language-based brain-machine interfaces in the future, especially for assisting aphasics in communicating more efficiently with fMRI signals.

What is the Link Between Mental Imagery and Sensory Sensitivity? Insights from Aphantasia

People with aphantasia have impoverished visual imagery so struggle to form mental pictures in the mind's eye. By testing people with and without aphantasia, we investigate the relationship between sensory imagery and sensory sensitivity (i.e., hyper- or hypo-reactivity to incoming signals through the sense organs). In Experiment 1 we first show that people with aphantasia report impaired imagery across multiple domains (e.g., olfactory, gustatory etc.) rather than simply vision. Importantly, we also show that imagery is related to sensory sensitivity: aphantasics reported not only lower imagery, but also lower sensory sensitivity. In Experiment 2, we showed a similar relationship between imagery and sensitivity in the general population. Finally, in Experiment 3 we found behavioural corroboration in a Pattern Glare Task, in which aphantasics experienced less visual discomfort and fewer visual distortions typically associated with sensory sensitivity. Our results suggest for the very first time that sensory imagery and sensory sensitivity are related, and that aphantasics are characterised by both lower imagery, and lower sensitivity. Our results also suggest that aphantasia (absence of visual imagery) may be more accurately defined as a subtype of a broader imagery deficit we name dysikonesia, in which weak or absent imagery occurs across multiple senses.