Object Representations in Human Visual Cortex Formed Through Temporal Integration of Dynamic Partial Shape Views

Evaluating integration of letter fragments through contrast and spatially targeted masking

Four experiments were conducted to gain a better understanding of the visual mechanisms related to how integration of partial shape cues provides for recognition of the full shape. In each experiment, letters formed as outline contours were displayed as a sequence of adjacent segments (fragments), each visible during a 17-ms time frame. The first experiment varied the contrast of the fragments. There were substantial individual differences in contrast sensitivity, so stimulus displays in the masking experiments that followed were calibrated to the sensitivity of each participant. Masks were displayed either as patterns that filled the entire screen (full field) or as successive strips that were sliced from the pattern, each strip lying across the location of the letter fragment that had been shown a moment before. Contrast of masks were varied to be lighter or darker than the letter fragments. Full-field masks, whether light or dark, provided relatively little impairment of recognition, as was the case for mask strips that were lighter than the letter fragments. However, dark strip masks proved to be very effective, with the degree of recognition impairment becoming larger as mask contrast was increased. A final experiment found the strip masks to be most effective when they overlapped the location where the letter fragments had been shown a moment before. They became progressively less effective with increased spatial separation from that location. Results are discussed with extensive reference to potential brain mechanisms for integrating shape cues.

The perceptual timescape: Perceptual history on the sub-second scale

There is a high-capacity store of brief time span (∼1000 ms) which information enters from perceptual processing, often called iconic memory or sensory memory. It is proposed that a main function of this store is to hold recent perceptual information in a temporally segregated representation, named the perceptual timescape. The perceptual timescape is a continually active representation of change and continuity over time that endows the perceived present with a perceived history. This is accomplished primarily by two kinds of time marking information: time distance information, which marks all items of information in the perceptual timescape according to how far in the past they occurred, and ordinal temporal information, which organises items of information in terms of their temporal order. Added to that is information about connectivity of perceptual objects over time. These kinds of information connect individual items over a brief span of time so as to represent change, persistence, and continuity over time. It is argued that there is a one-way street of information flow from perceptual processing either to the perceived present or directly into the perceptual timescape, and thence to working memory. Consistent with that, the information structure of the perceptual timescape supports postdictive reinterpretations of recent perceptual information. Temporal integration on a time scale of hundreds of milliseconds takes place in perceptual processing and does not draw on information in the perceptual timescape, which is concerned with temporal segregation, not integration.

Hierarchical Bayesian Causality Network to Extract High-Level Semantic Information in Visual Cortex

Functional MRI (fMRI) is a brain signal with high spatial resolution, and visual cognitive processes and semantic information in the brain can be represented and obtained through fMRI. In this paper, we design single-graphic and matched/unmatched double-graphic visual stimulus experiments and collect 12 subjects' fMRI data to explore the brain's visual perception processes. In the double-graphic stimulus experiment, we focus on the high-level semantic information as "matching", and remove tail-to-tail conjunction by designing a model to screen the matching-related voxels. Then, we perform Bayesian causal learning between fMRI voxels based on the transfer entropy, establish a hierarchical Bayesian causal network (HBcausalNet) of the visual cortex, and use the model for visual stimulus image reconstruction. HBcausalNet achieves an average accuracy of 70.57% and 53.70% in single- and double-graphic stimulus image reconstruction tasks, respectively, higher than HcorrNet and HcasaulNet. The results show that the matching-related voxel screening and causality analysis method in this paper can extract the "matching" information in fMRI, obtain a direct causal relationship between matching information and fMRI, and explore the causal inference process in the brain. It suggests that our model can effectively extract high-level semantic information in brain signals and model effective connections and visual perception processes in the visual cortex of the brain.

Orientation representation in human visual cortices: contributions of non-visual information and action-related process

Orientation processing in the human brain plays a crucial role in guiding grasping actions toward an object. Remarkably, despite the absence of visual input, the human visual cortex can still process orientation information. Instead of visual input, non-visual information, including tactile and proprioceptive sensory input from the hand and arm, as well as feedback from action-related processes, may contribute to orientation processing. However, the precise mechanisms by which the visual cortices process orientation information in the context of non-visual sensory input and action-related processes remain to be elucidated. Thus, our study examined the orientation representation within the visual cortices by analyzing the blood-oxygenation-level-dependent (BOLD) signals under four action conditions: direct grasp (DG), air grasp (AG), non-grasp (NG), and uninformed grasp (UG). The images of the cylindrical object were shown at +45° or - 45° orientations, corresponding to those of the real object to be grasped with the whole-hand gesture. Participants judged their orientation under all conditions. Grasping was performed without online visual feedback of the hand and object. The purpose of this design was to investigate the visual areas under conditions involving tactile feedback, proprioception, and action-related processes. To address this, a multivariate pattern analysis was used to examine the differences among the cortical patterns of the four action conditions in orientation representation by classification. Overall, significant decoding accuracy over chance level was discovered for the DG; however, during AG, only the early visual areas showed significant accuracy, suggesting that the object's tactile feedback influences the orientation process in higher visual areas. The NG showed no statistical significance in any area, indicating that without the grasping action, visual input does not contribute to cortical pattern representation. Interestingly, only the dorsal and ventral divisions of the third visual area (V3d and V3v) showed significant decoding accuracy during the UG despite the absence of visual instructions, suggesting that the orientation representation was derived from action-related processes in V3d and visual recognition of object visualization in V3v. The processing of orientation information during non-visually guided grasping of objects relies on other non-visual sources and is specifically divided by the purpose of action or recognition.

Recognition of letters displayed as successive contour fragments

Shapes can be displayed as parts but perceived as a whole through feedforward and feedback mechanisms in the visual system, though the exact spatiotemporal relationships for this process are still unclear. Our experiments examined the integration of letter fragments that were displayed as a rapid sequence. We examined the effects of timing and masking on integration, hypothesizing that increasing the timing interval between frames would impair recognition by disrupting contour linkage. We further used different mask types, a full-field pattern mask and a smaller strip mask, to examine the effects of global vs local masking on integration. We found that varying mask types and contrast produced a greater decline in recognition than was found when persistence or mask density was manipulated. The study supports prior work on letter recognition and provides greater insight into the spatiotemporal factors that contribute to the identification of shapes.

On the Spatiotemporal Nature of Vision, as Revealed by Covered Bridges and Puddles: A Dispatch from Vermont

Retinal painting, anorthoscopic perception and amodal completion are terms to describe visual phenomena that highlight the spatiotemporal integrative mechanisms that underlie primate vision. Although commonly studied using simplified lab-friendly stimuli presented on a computer screen, this is a report of observations made in a novel real-world context that highlight the rich contributions the mechanisms underlying these phenomena make to naturalistic vision.

The chronometry of symmetry detection in the lateral occipital (LO) cortex

The lateral occipital cortex (LO) has been shown to code the presence of both vertical and horizontal visual symmetry in dot patterns. However, the specific time window at which LO is causally involved in symmetry encoding has not been investigated. This was assessed using a chronometric transcranial magnetic stimulation (TMS) approach. Participants were presented with a series of dot configurations and instructed to judge whether they were symmetric along the vertical axis or not while receiving a double pulse of TMS over either the right LO (rLO) or the vertex (baseline) at different time windows (ranging from 50 ms to 290 ms from stimulus onset). We found that TMS delivered over the rLO significantly decreased participants' accuracy in discriminating symmetric from non-symmetric patterns when TMS was applied between 130 ms and 250 ms from stimulus onset, suggesting that LO is causally involved in symmetry perception within this time window. These findings confirm and extend prior neuroimaging and ERP evidence by demonstrating not only that LO is causally involved in symmetry encoding but also that its contribution occurs in a relatively large temporal window, at least in tasks requiring fast discrimination of mirror symmetry in briefly (75 ms) presented patterns as in our study.

Spinning objects and partial occlusion: Smart neural responses to symmetry

In humans, extrastriate visual areas are strongly activated by symmetry. However, perfect symmetry is rare in natural visual images. Recent findings showed that when parts of a symmetric shape are presented at different points in time the process relies on a perceptual memory buffer. Does this temporal integration need a retinotopic reference frame? For the first time we tested integration of parts both in the temporal and spatial domain, using a non-retinotopic frame of reference. In Experiment 1, an irregular polygonal shape (either symmetric or asymmetric) was partly occluded by a rectangle for 500 ms (T1). The rectangle moved to the opposite side to reveal the other half of the shape, whilst occluding the previously visible half (T2). The reference frame for the object was static: the two parts stimulated retinotopically corresponding receptive fields (revealed over time). A symmetry-specific ERP response from ~300 ms after T2 was observed. In Experiment 2 dynamic occlusion was combined with an additional step at T2: the new half-shape and occluder were rotated by 90°. Therefore, there was a moving frame of reference and the retinal correspondence between the two parts was disrupted. A weaker but significant symmetry-specific response was recorded. This result extends previous findings: global symmetry representation can be achieved in extrastriate areas non-retinotopically, through integration in both temporal and spatial domain.

Moving a Shape behind a Slit: Partial Shape Representations in Inferior Temporal Cortex

Current models of object recognition are based on spatial representations build from object features that are simultaneously present in the retinal image. However, one can recognize an object when it moves behind a static occlude, and only a small fragment of its shape is visible through a slit at a given moment in time. Such anorthoscopic perception requires spatiotemporal integration of the successively presented shape parts during slit-viewing. Human fMRI studies suggested that ventral visual stream areas represent whole shapes formed through temporal integration during anorthoscopic perception. To examine the time course of shape-selective responses during slit-viewing, we recorded the responses of single inferior temporal (IT) neurons of rhesus monkeys to moving shapes that were only partially visible through a static narrow slit. The IT neurons signaled shape identity by their response when that was cumulated across the duration of the shape presentation. Their shape preference during slit-viewing equaled that for static, whole-shape presentations. However, when analyzing their responses at a finer time scale, we showed that the IT neurons responded to particular shape fragments that were revealed by the slit. We found no evidence for temporal integration of slit-views that result in a whole-shape representation, even when the monkey was matching slit-views of a shape to static whole-shape presentations. These data suggest that, although the temporally integrated response of macaque IT neurons can signal shape identity in slit-viewing conditions, the spatiotemporal integration needed for the formation of a whole-shape percept occurs in other areas, perhaps downstream to IT.SIGNIFICANCE STATEMENT One recognizes an object when it moves behind a static occluder and only a small fragment of its shape is visible through a static slit at a given moment in time. Such anorthoscopic perception requires spatiotemporal integration of the successively presented partial shape parts. Human fMRI studies suggested that ventral visual stream areas represent shapes formed through temporal integration. We recorded the responses of inferior temporal (IT) cortical neurons of macaques during slit-viewing conditions. Although the temporally summated response of macaque IT neurons could signal shape identity under slit-viewing conditions, we found no evidence for a whole-shape representation using analyses at a finer time scale. Thus, the spatiotemporal integration needed for anorthoscopic perception does not occur within IT.

Attention control processes that prioritise task execution may come at the expense of incidental memory encoding

Attention underpins episodic memory encoding by gating information processing. However, it is unclear how different forms of attention affect encoding. Using fMRI, we implemented a novel task that separates top-down and bottom-up attention (TDA; BUA) to test how these forms of attention influence encoding. Twenty-seven subjects carried out a scanned incidental encoding task that required semantic categorisation of stimuli. Trials either required visual search (TDA) to locate a target, or the target blinked and captured attention (BUA). After a retention period, subjects performed a surprise recognition test. Univariate analyses showed that ventral visual regions and right hippocampus indexed encoding success. Psychophysiological interaction analyses showed that, during TDA, there was increased coupling between dorsal parietal cortex and fusiform gyrus with encoding failure, and between lateral occipital cortex and fusiform gyrus with encoding success. No significant connectivity modulations were observed during BUA. We propose that increased TDA to objects in space is mediated by parietal cortex and negatively impacts encoding. Also, increases in connectivity within ventral visual cortex index the integration of stimulus features, promoting encoding. Finally, the influences of attention on encoding likely depend on task demands: as cognitive control increases, task execution is emphasised at the expense of memory encoding.

Topics and trends in artificial intelligence assisted human brain research

Artificial intelligence (AI) assisted human brain research is a dynamic interdisciplinary field with great interest, rich literature, and huge diversity. The diversity in research topics and technologies keeps increasing along with the tremendous growth in application scope of AI-assisted human brain research. A comprehensive understanding of this field is necessary to assess research efficacy, (re)allocate research resources, and conduct collaborations. This paper combines the structural topic modeling (STM) with the bibliometric analysis to automatically identify prominent research topics from the large-scale, unstructured text of AI-assisted human brain research publications in the past decade. Analyses on topical trends, correlations, and clusters reveal distinct developmental trends of these topics, promising research orientations, and diverse topical distributions in influential countries/regions and research institutes. These findings help better understand scientific and technological AI-assisted human brain research, provide insightful guidance for resource (re)allocation, and promote effective international collaborations.

Temporal integration in the perception and discrimination of solid shape

Two experiments evaluated the importance of temporal integration for the perception and discrimination of solid object shape. In Experiment 1, observers anorthoscopically viewed moving or stationary cast shadows of naturally shaped solid objects (bell peppers, Capsicum annuum) through narrow (4-mm wide) slits. At any given moment, observers could only see a very small portion of the overall object shape (generally less than 10%). The results showed that the observers' discrimination performance for the moving cast shadows was much higher than that obtained for the stationary shadows, demonstrating the ability to temporally integrate the piecemeal momentary information about shape that was available through the narrow apertures. In a second experiment, estimates of the strength of the observers' impressions of solid shapes rotating in depth were obtained as well as discrimination accuracies; perceptions of the original moving condition were compared with a new condition where the frames of the apparent motion sequences depicting solid objects in continuous motion (behind the slits) were randomly scrambled. The observers perceived the anorthoscopic displays as depicting solid objects rotating in depth, but only in the continuous motion condition. Interestingly, the discrimination performance in the scrambled condition remained relatively high-observers were still able to integrate information across the multiple scrambled frames in order to produce discrimination performance that was significantly higher than that obtained in the stationary shadow condition. This study was the first to thoroughly evaluate whether and to what extent human observers can effectively discriminate and perceive solid object shape anorthoscopically.

Neuronal correlates of full and partial visual conscious perception

Stimuli may induce only partial consciousness-an intermediate between null and full consciousness-where the presence but not identity of an object can be reported. The differences in the neuronal basis of full and partial consciousness are poorly understood. We investigated if evoked and oscillatory activity could dissociate full from partial conscious perception. We recorded human cortical activity with magnetoencephalography (MEG) during a visual perception task in which stimulus could be either partially or fully perceived. Partial consciousness was associated with an early increase in evoked activity and theta/low-alpha-band oscillations while full consciousness was also associated with late evoked activity and beta-band oscillations. Full from partial consciousness was dissociated by stronger evoked activity and late increase in theta oscillations that were localized to higher-order visual regions and posterior parietal and prefrontal cortices. Our results reveal both evoked activity and theta oscillations dissociate partial and full consciousness.