The right posterior parietal cortex mediates spatial reorienting of attentional choice bias.

Attention facilitates behavior by enhancing perceptual sensitivity (sensory processing) and choice bias (decisional weighting) for attended information. Whether distinct neural substrates mediate these distinct components of attention remains unknown. We investigate the causal role of key nodes of the right posterior parietal cortex (rPPC) in the forebrain attention network in sensitivity versus bias control. Two groups of participants performed a cued attention task while we applied either inhibitory, repetitive transcranial magnetic stimulation (n = 28) or 40 Hz transcranial alternating current stimulation (n = 26) to the dorsal rPPC. We show that rPPC stimulation - with either modality - impairs task performance by selectively altering attentional modulation of bias but not sensitivity. Specifically, participants' bias toward the uncued, but not the cued, location reduced significantly following rPPC stimulation - an effect that was consistent across both neurostimulation cohorts. In sum, the dorsal rPPC causally mediates the reorienting of choice bias, one particular component of visual spatial attention.

Dissociable components of attention exhibit distinct neuronal signatures in primate visual cortex.

Attention can be deployed in multiple forms and facilitates behavior by influencing perceptual sensitivity and choice bias. Attention is also associated with a myriad of changes in sensory neural activity. Yet, the relationship between the behavioral components of attention and the accompanying changes in neural activity remains largely unresolved. We examined this relationship by quantifying sensitivity and bias in monkeys performing a task that dissociated eye movement responses from the focus of covert attention. Unexpectedly, bias, not sensitivity, increased at the focus of covert attention, whereas sensitivity increased at the location of planned eye movements. Furthermore, neuronal activity within visual area V4 varied robustly with bias, but not sensitivity, at the focus of covert attention. In contrast, correlated variability between neuronal pairs was lowest at the location of planned eye movements, and varied with sensitivity, but not bias. Thus, dissociable behavioral components of attention exhibit distinct neuronal signatures within the visual cortex.

Presaccadic attention does not facilitate the detection of changes in the visual field.

Planning a rapid eye movement (saccade) changes how we perceive our visual world. Even before we move the eyes visual discrimination sensitivity improves at the impending target of eye movements, a phenomenon termed "presaccadic attention." Yet, it is unknown if such presaccadic selection merely affects perceptual sensitivity, or also affects downstream decisional processes, such as choice bias. We report a surprising lack of presaccadic perceptual benefits in a common, everyday setting-detection of changes in the visual field. Despite the lack of sensitivity benefits, choice bias for reporting changes increased reliably for the saccade target. With independent follow-up experiments, we show that presaccadic change detection is rendered more challenging because percepts at the saccade target location are biased toward, and more precise for, only the most recent of two successive stimuli. With a Bayesian model, we show how such perceptual and choice biases are crucial to explain the effects of saccade plans on change detection performance. In sum, visual change detection sensitivity does not improve presaccadically, a result that is readily explained by teasing apart distinct components of presaccadic selection. The findings may have critical implications for real-world scenarios, like driving, that require rapid gaze shifts in dynamically changing environments.

Tracking momentary fluctuations in human attention with a cognitive brain-machine interface.

Selective attention produces systematic effects on neural states. It is unclear whether, conversely, momentary fluctuations in neural states have behavioral significance for attention. We investigated this question in the human brain with a cognitive brain-machine interface (cBMI) for tracking electrophysiological steady-state visually evoked potentials (SSVEPs) in real-time. Discrimination accuracy (d') was significantly higher when target stimuli were triggered at high, versus low, SSVEP power states. Target and distractor SSVEP power was uncorrelated across the hemifields, and target d' was unaffected by distractor SSVEP power states. Next, we trained participants on an auditory neurofeedback paradigm to generate biased, cross-hemispheric competitive interactions between target and distractor SSVEPs. The strongest behavioral effects emerged when competitive SSVEP dynamics unfolded at a timescale corresponding to the deployment of endogenous attention. In sum, SSVEP power dynamics provide a reliable readout of attentional state, a result with critical implications for tracking and training human attention.

A Midbrain Inspired Recurrent Neural Network Model for Robust Change Detection.

We present a biologically inspired recurrent neural network (RNN) that efficiently detects changes in natural images. The model features sparse, topographic connectivity (st-RNN), closely modeled on the circuit architecture of a "midbrain attention network." We deployed the st-RNN in a challenging change blindness task, in which changes must be detected in a discontinuous sequence of images. Compared with a conventional RNN, the st-RNN learned 9x faster and achieved state-of-the-art performance with 15x fewer connections. An analysis of low-dimensional dynamics revealed putative circuit mechanisms, including a critical role for a global inhibitory (GI) motif, for successful change detection. The model reproduced key experimental phenomena, including midbrain neurons' sensitivity to dynamic stimuli, neural signatures of stimulus competition, as well as hallmark behavioral effects of midbrain microstimulation. Finally, the model accurately predicted human gaze fixations in a change blindness experiment, surpassing state-of-the-art saliency-based methods. The st-RNN provides a novel deep learning model for linking neural computations underlying change detection with psychophysical mechanisms.SIGNIFICANCE STATEMENT For adaptive survival, our brains must be able to accurately and rapidly detect changing aspects of our visual world. We present a novel deep learning model, a sparse, topographic recurrent neural network (st-RNN), that mimics the neuroanatomy of an evolutionarily conserved "midbrain attention network." The st-RNN achieved robust change detection in challenging change blindness tasks, outperforming conventional RNN architectures. The model also reproduced hallmark experimental phenomena, both neural and behavioral, reported in seminal midbrain studies. Lastly, the st-RNN outperformed state-of-the-art models at predicting human gaze fixations in a laboratory change blindness experiment. Our deep learning model may provide important clues about key mechanisms by which the brain efficiently detects changes.

GPU-accelerated connectome discovery at scale.

Diffusion magnetic resonance imaging and tractography enable the estimation of anatomical connectivity in the human brain, in vivo. Yet, without ground-truth validation, different tractography algorithms can yield widely varying connectivity estimates. Although streamline pruning techniques mitigate this challenge, slow compute times preclude their use in big-data applications. We present 'Regularized, Accelerated, Linear Fascicle Evaluation' (ReAl-LiFE), a GPU-based implementation of a state-of-the-art streamline pruning algorithm (LiFE), which achieves >100× speedups over previous CPU-based implementations. Leveraging these speedups, we overcome key limitations with LiFE's algorithm to generate sparser and more accurate connectomes. We showcase ReAl-LiFE's ability to estimate connections with superlative test-retest reliability, while outperforming competing approaches. Moreover, we predicted inter-individual variations in multiple cognitive scores with ReAl-LiFE connectome features. We propose ReAl-LiFE as a timely tool, surpassing the state of the art, for accurate discovery of individualized brain connectomes at scale. Finally, our GPU-accelerated implementation of a popular non-negative least-squares optimization algorithm is widely applicable to many real-world problems.

Neurally-constrained modeling of human gaze strategies in a change blindness task.

Despite possessing the capacity for selective attention, we often fail to notice the obvious. We investigated participants' (n = 39) failures to detect salient changes in a change blindness experiment. Surprisingly, change detection success varied by over two-fold across participants. These variations could not be readily explained by differences in scan paths or fixated visual features. Yet, two simple gaze metrics-mean duration of fixations and the variance of saccade amplitudes-systematically predicted change detection success. We explored the mechanistic underpinnings of these results with a neurally-constrained model based on the Bayesian framework of sequential probability ratio testing, with a posterior odds-ratio rule for shifting gaze. The model's gaze strategies and success rates closely mimicked human data. Moreover, the model outperformed a state-of-the-art deep neural network (DeepGaze II) with predicting human gaze patterns in this change blindness task. Our mechanistic model reveals putative rational observer search strategies for change detection during change blindness, with critical real-world implications.

Decoding Task-Specific Cognitive States with Slow, Directed Functional Networks in the Human Brain.

Flexible functional interactions among brain regions mediate critical cognitive functions. Such interactions can be measured using functional magnetic resonance imaging (fMRI) data either with instantaneous (zero-lag) or lag-based (time-lagged) functional connectivity. Because the fMRI hemodynamic response is slow, and is sampled at a timescale (seconds) several orders of magnitude slower than the underlying neural dynamics (milliseconds), simulation studies have shown that lag-based fMRI functional connectivity, measured with approaches like Granger-Geweke causality (GC), provides spurious and unreliable estimates of underlying neural interactions. Experimental verification of this claim is challenging because neural ground truth connectivity is often unavailable concurrently with fMRI recordings. Here we demonstrate that, despite these widely held caveats, GC networks estimated from fMRI recordings contain useful information for classifying task-specific cognitive states. We estimated instantaneous and lag-based GC functional connectivity networks using fMRI data from 1000 participants (Human Connectome Project database). A linear classifier, trained on either instantaneous or lag-based GC, reliably discriminated among seven different task and resting brain states, with >80% cross-validation accuracy. With network simulations, we demonstrate that instantaneous and lag-based GC exploited interactions at fast and slow timescales, respectively, to achieve robust classification. With human fMRI data, instantaneous and lag-based GC identified complementary, task-core networks. Finally, variations in GC connectivity explained inter-individual variations in a variety of cognitive scores. Our findings show that instantaneous and lag-based methods reveal complementary aspects of functional connectivity in the brain, and suggest that slow, directed functional interactions, estimated with fMRI, may provide useful markers of behaviorally relevant cognitive states.

Subcortical connectivity correlates selectively with attention's effects on spatial choice bias.

Neural mechanisms of attention are extensively studied in the neocortex; comparatively little is known about how subcortical regions contribute to attention. The superior colliculus (SC) is an evolutionarily conserved, subcortical (midbrain) structure that has been implicated in controlling visuospatial attention. Yet how the SC contributes mechanistically to attention remains unknown. We investigated the role of the SC in attention, combining model-based psychophysics, diffusion imaging, and tractography in human participants. Specifically, we asked whether the SC contributes to enhancing sensitivity (d') to attended information, or whether it contributes to biasing choices (criteria) in favor of attended information. We tested human participants on a multialternative change detection task, with endogenous spatial cueing, and quantified sensitivity and bias with a recently developed multidimensional signal detection model (m-ADC model). At baseline, sensitivity and bias exhibited complementary patterns of asymmetries across the visual hemifields: While sensitivity was consistently higher for detecting changes in the left hemifield, bias was higher for reporting changes in the right hemifield. Remarkably, white matter connectivity of the SC with the neocortex mirrored this pattern of asymmetries. Specifically, the asymmetry in SC-cortex connectivity correlated with the asymmetry in choice bias, but not in sensitivity. In addition, SC-cortex connectivity strength could predict cueing-induced modulation of bias, but not of sensitivity, across individuals. In summary, the SC may be a key node in an evolutionarily conserved network for controlling choice bias during visuospatial attention.

Dissociable sensitivity and bias mechanisms mediate behavioral effects of exogenous attention.

Attention can be directed endogenously, based on task-relevant goals, or captured exogenously, by salient stimuli. While recent studies have shown that endogenous attention can facilitate behavior through dissociable sensitivity (sensory) and choice bias (decisional) mechanisms, it is unknown if exogenous attention also operates through dissociable sensitivity and bias mechanisms. We tested human participants on a multialternative change detection task with exogenous attention cues, which preceded or followed change events in close temporal proximity. Analyzing participants' behavior with a multidimensional signal detection model revealed clear dissociations between exogenous cueing effects on sensitivity and bias. While sensitivity was, overall, lower at the cued location compared to other locations, bias was highest at the cued location. With an appropriately designed post-cue control condition, we discovered that the attentional effect of exogenous pre-cueing was to enhance sensitivity proximal to the cue. In contrast, exogenous attention enhanced bias even for distal stimuli in the cued hemifield. Reaction time effects of exogenous cueing could be parsimoniously explained with a diffusion-decision model, in which drift rate was determined by independent contributions from sensitivity and bias at each location. The results suggest a mechanistic schema of how exogenous attention engages dissociable sensitivity and bias mechanisms to shape behavior.

ReAl-LiFE: Accelerating the Discovery of Individualized Brain Connectomes on GPUs.

Diffusion imaging and tractography enable mapping structural connections in the human brain, in-vivo. Linear Fascicle Evaluation (LiFE) is a state-of-the-art approach for pruning spurious connections in the estimated structural connectome, by optimizing its fit to the measured diffusion data. Yet, LiFE imposes heavy demands on computing time, precluding its use in analyses of large connectome databases. Here, we introduce a GPU-based implementation of LiFE that achieves 50-100x speedups over conventional CPU-based implementations for connectome sizes of up to several million fibers. Briefly, the algorithm accelerates generalized matrix multiplications on a compressed tensor through efficient GPU kernels, while ensuring favorable memory access patterns. Leveraging these speedups, we advance LiFE's algorithm by imposing a regularization constraint on estimated fiber weights during connectome pruning. Our regularized, accelerated, LiFE algorithm ("ReAl-LiFE") estimates sparser connectomes that also provide more accurate fits to the underlying diffusion signal. We demonstrate the utility of our approach by classifying pathological signatures of structural connectivity in patients with Alzheimer's Disease (AD). We estimated million fiber whole-brain connectomes, followed by pruning with ReAl-LiFE, for 90 individuals (45 AD patients and 45 healthy controls). Linear classifiers, based on support vector machines, achieved over 80% accuracy in classifying AD patients from healthy controls based on their ReAl-LiFE pruned structural connectomes alone. Moreover, classification based on the ReAl-LiFE pruned connectome outperformed both the unpruned connectome, as well as the LiFE pruned connectome, in terms of accuracy. We propose our GPU-accelerated approach as a widely relevant tool for non-negative least squares optimization, across many domains.

Mapping distinct timescales of functional interactions among brain networks.

Brain processes occur at various timescales, ranging from milliseconds (neurons) to minutes and hours (behavior). Characterizing functional coupling among brain regions at these diverse timescales is key to understanding how the brain produces behavior. Here, we apply instantaneous and lag-based measures of conditional linear dependence, based on Granger-Geweke causality (GC), to infer network connections at distinct timescales from functional magnetic resonance imaging (fMRI) data. Due to the slow sampling rate of fMRI, it is widely held that GC produces spurious and unreliable estimates of functional connectivity when applied to fMRI data. We challenge this claim with simulations and a novel machine learning approach. First, we show, with simulated fMRI data, that instantaneous and lag-based GC identify distinct timescales and complementary patterns of functional connectivity. Next, we analyze fMRI scans from 500 subjects and show that a linear classifier trained on either instantaneous or lag-based GC connectivity reliably distinguishes task versus rest brain states, with ~80-85% cross-validation accuracy. Importantly, instantaneous and lag-based GC exploit markedly different spatial and temporal patterns of connectivity to achieve robust classification. Our approach enables identifying functionally connected networks that operate at distinct timescales in the brain.

Sensory and decisional components of endogenous attention are dissociable.

Endogenous cueing of attention enhances sensory processing of the attended stimulus (perceptual sensitivity) and prioritizes information from the attended location for guiding behavioral decisions (spatial choice bias). Here, we test whether sensitivity and bias effects of endogenous spatial attention are under the control of common or distinct mechanisms. Human observers performed a multialternative visuospatial attention task with probabilistic spatial cues. Observers' behavioral choices were analyzed with a recently developed multidimensional signal detection model (the m-ADC model). The model effectively decoupled the effects of spatial cueing on sensitivity from those on spatial bias and revealed striking dissociations between them. Sensitivity was highest at the cued location and not significantly different among uncued locations, suggesting a spotlight-like allocation of sensory resources at the cued location. On the other hand, bias varied systematically with cue validity, suggesting a graded allocation of decisional priority across locations. Cueing-induced modulations of sensitivity and bias were uncorrelated within and across subjects. Bias, but not sensitivity, correlated with key metrics of prioritized decision-making, including reaction times and decision optimality indices. In addition, we developed a novel metric, differential risk curvature, for distinguishing bias effects of attention from those of signal expectation. Differential risk curvature correlated selectively with m-ADC model estimates of bias but not with estimates of sensitivity. Our results reveal dissociable effects of endogenous attention on perceptual sensitivity and choice bias in a multialternative choice task and motivate the search for the distinct neural correlates of each.NEW & NOTEWORTHY Attention is often studied as a unitary phenomenon. Yet, attention can both enhance the perception of important stimuli (sensitivity) and prioritize such stimuli for decision-making (bias). Employing a multialternative spatial attention task with probabilistic cueing, we show that attention affects sensitivity and bias through dissociable mechanisms. Specifically, the effects on sensitivity alone match the notion of an attentional "spotlight." Our behavioral model enables quantifying component processes of attention, and identifying their respective neural correlates.

Unraveling Causal Mechanisms of Top-Down and Bottom-Up Visuospatial Attention with Non-invasive Brain Stimulation.

Attention is a process of selection that allows us to intelligently navigate the abundance of information in our world. Attention can be either directed voluntarily based on internal goals-"top-down" or goal-directed attention-or captured automatically, by salient stimuli-"bottom-up" or stimulus-driven attention. Do these two modes of attention control arise from same or different brain circuits? Do they share similar or distinct neural mechanisms? In this review, we explore this dichotomy between the neural bases of top-down and bottom-up attention control, with a special emphasis on insights gained from non-invasive neurostimulation techniques, specifically, transcranial magnetic stimulation (TMS). TMS enables spatially focal and temporally precise manipulation of brain activity. We explore a significant literature devoted to investigating the role of fronto-parietal brain regions in top-down and bottom-up attention with TMS, and highlight key areas of convergence and debate. We also discuss recent advances in combinatorial paradigms that combine TMS with other imaging modalities, such as functional magnetic resonance imaging or electroencephalography. These paradigms are beginning to bridge essential gaps in our understanding of the neural pathways by which TMS affects behavior, and will prove invaluable for unraveling mechanisms of attention control, both in health and in disease.

Infra-slow brain dynamics as a marker for cognitive function and decline.

Functional magnetic resonance imaging (fMRI) enables measuring human brain activity, in vivo. Yet, the fMRI hemodynamic response unfolds over very slow timescales (<0.1-1 Hz), orders of magnitude slower than millisecond timescales of neural spiking. It is unclear, therefore, if slow dynamics as measured with fMRI are relevant for cognitive function. We investigated this question with a novel application of Gaussian Process Factor Analysis (GPFA) and machine learning to fMRI data. We analyzed slowly sampled (1.4 Hz) fMRI data from 1000 healthy human participants (Human Connectome Project database), and applied GPFA to reduce dimensionality and extract smooth latent dynamics. GPFA dimensions with slow (<1 Hz) characteristic timescales identified, with high accuracy (>95%), the specific task that each subject was performing inside the fMRI scanner. Moreover, functional connectivity between slow GPFA latents accurately predicted inter-individual differences in behavioral scores across a range of cognitive tasks. Finally, infra-slow (<0.1 Hz) latent dynamics predicted CDR (Clinical Dementia Rating) scores of individual patients, and identified patients with mild cognitive impairment (MCI) who would progress to develop Alzheimer's dementia (AD). Slow and infra-slow brain dynamics may be relevant for understanding the neural basis of cognitive function, in health and disease.

Space-Specific Deficits in Visual Orientation Discrimination Caused by Lesions in the Midbrain Stimulus Selection Network.

Perceptual decisions require both analysis of sensory information and selective routing of relevant information to decision networks. This study explores the contribution of a midbrain network to visual perception in chickens. Analysis of visual orientation information in birds takes place in the forebrain sensory area called the Wulst, as it does in the primary visual cortex (V1) of mammals. In contrast, the midbrain, which receives parallel retinal input, encodes orientation poorly, if at all. We discovered, however, that small electrolytic lesions in the midbrain severely impair a chicken's ability to discriminate orientations. Focal lesions were placed in the optic tectum (OT) and in the nucleus isthmi pars parvocellularis (Ipc)-key nodes in the midbrain stimulus selection network-in chickens trained to perform an orientation discrimination task. A lesion in the OT caused a severe impairment in orientation discrimination specifically for targets at the location in space represented by the lesioned location. Distracting stimuli increased the deficit. A lesion in the Ipc produced similar but more transient effects. We discuss the possibilities that performance deficits were caused by interference with orientation information processing (sensory deficit) versus with the routing of information in the forebrain (agnosia). The data support the proposal that the OT transmits a space-specific signal that is required to gate orientation information from the Wulst into networks that mediate behavioral decisions, analogous to the role of ascending signals from the superior colliculus (SC) in monkeys. Furthermore, our results indicate a critical role for the cholinergic Ipc in this gating process.

Does the Superior Colliculus Control Perceptual Sensitivity or Choice Bias during Attention? Evidence from a Multialternative Decision Framework.

Distinct networks in the forebrain and the midbrain coordinate to control spatial attention. The critical involvement of the superior colliculus (SC)-the central structure in the midbrain network-in visuospatial attention has been shown by four seminal, published studies in monkeys (Macaca mulatta) performing multialternative tasks. However, due to the lack of a mechanistic framework for interpreting behavioral data in such tasks, the nature of the SC's contribution to attention remains unclear. Here we present and validate a novel decision framework for analyzing behavioral data in multialternative attention tasks. We apply this framework to re-examine the behavioral evidence from these published studies. Our model is a multidimensional extension to signal detection theory that distinguishes between two major classes of attentional mechanisms: those that alter the quality of sensory information or "sensitivity," and those that alter the selective gating of sensory information or "choice bias." Model-based simulations and model-based analyses of data from these published studies revealed a converging pattern of results that indicated that choice-bias changes, rather than sensitivity changes, were the primary outcome of SC manipulation. Our results suggest that the SC contributes to attentional performance predominantly by generating a spatial choice bias for stimuli at a selected location, and that this bias operates downstream of forebrain mechanisms that enhance sensitivity. The findings lead to a testable mechanistic framework of how the midbrain and forebrain networks interact to control spatial attention. SIGNIFICANCE STATEMENT: Attention involves the selection of the most relevant information for differential sensory processing and decision making. While the mechanisms by which attention alters sensory encoding (sensitivity control) are well studied, the mechanisms by which attention alters decisional weighting of sensory evidence (choice-bias control) are poorly understood. Here, we introduce a model of multialternative decision making that distinguishes bias from sensitivity effects in attention tasks. With our model, we simulate experimental data from four seminal studies that microstimulated or inactivated a key attention-related midbrain structure, the superior colliculus (SC). We demonstrate that the experimental effects of SC manipulation are entirely consistent with the SC controlling attention by changing choice bias, thereby shedding new light on how the brain mediates attention.

Gamma oscillations in the midbrain spatial attention network: linking circuits to function.

Gamma-band (25-140Hz) oscillations are ubiquitous in mammalian forebrain structures involved in sensory processing, attention, learning and memory. The optic tectum (OT) is the central structure in a midbrain network that participates critically in controlling spatial attention. In this review, we summarize recent advances in characterizing a neural circuit in this midbrain network that generates large amplitude, space-specific, gamma oscillations in the avian OT, both in vivo and in vitro. We describe key physiological and pharmacological mechanisms that produce and regulate the structure of these oscillations. The extensive similarities between midbrain gamma oscillations in birds and those in the neocortex and hippocampus of mammals, offer important insights into the functional significance of a midbrain gamma oscillatory code.

Selective disinhibition: A unified neural mechanism for predictive and post hoc attentional selection.

The natural world presents us with a rich and ever-changing sensory landscape containing diverse stimuli that constantly compete for representation in the brain. When the brain selects a stimulus as the highest priority for attention, it differentially enhances the representation of the selected, "target" stimulus and suppresses the processing of other, distracting stimuli. A stimulus may be selected for attention while it is still present in the visual scene (predictive selection) or after it has vanished (post hoc selection). We present a biologically inspired computational model that accounts for the prioritized processing of information about targets that are selected for attention either predictively or post hoc. Central to the model is the neurobiological mechanism of "selective disinhibition" - the selective suppression of inhibition of the representation of the target stimulus. We demonstrate that this mechanism explains major neurophysiological hallmarks of selective attention, including multiplicative neural gain, increased inter-trial reliability (decreased variability), and reduced noise correlations. The same mechanism also reproduces key behavioral hallmarks associated with target-distracter interactions. Selective disinhibition exhibits several distinguishing and advantageous features over alternative mechanisms for implementing target selection, and is capable of explaining the effects of selective attention over a broad range of real-world conditions, involving both predictive and post hoc biasing of sensory competition and decisions.