Forgetting in visual working memory: Internal noise explains decay of feature representations.

The precision of visual working memory (VWM) representations decreases as time passes. It is often assumed that VWM decay is random and caused by internal noise accumulation. However, forgetting in VWM could occur systematically, such that some features deteriorate more rapidly than others. There exist only a few studies testing these two models of forgetting, with conflicting results. Here, decay of features in VWM was thoroughly tested using signal detection theory methods: psychophysical classification images, internal noise estimation, and receiver operant characteristic (ROC). A modified same-different memory task was employed with two retention times (500 and 4000 ms). Experiment 1 investigated VWM decay using a compound grating memory task, and Experiment 2 tested shape memory using radial frequency patterns. Memory performance dropped some 15% with increasing retention time in both experiments. Interestingly, classification images showed virtually indistinguishable weighting of stimulus features at both retention times, suggesting that VWM decay is not feature specific. Instead, we found a 77% increase in stimulus-independent internal noise at the longer retention time. Finally, the slope of the ROC curve plotted as z-scores was shallower at the longer retention time, indicating that the amount of stimulus-independent internal noise increased. Together these findings provide strong support for the idea that VWM decay does not result from a systematic loss of some stimulus features but instead is caused by uniformly increasing random internal noise.

Template optimization and transfer in perceptual learning.

We studied how learning changes the processing of a low-level Gabor stimulus, using a classification-image method (psychophysical reverse correlation) and a task where observers discriminated between slight differences in the phase (relative alignment) of a target Gabor in visual noise. The method estimates the internal "template" that describes how the visual system weights the input information for decisions. One popular idea has been that learning makes the template more like an ideal Bayesian weighting; however, the evidence has been indirect. We used a new regression technique to directly estimate the template weight change and to test whether the direction of reweighting is significantly different from an optimal learning strategy. The subjects trained the task for six daily sessions, and we tested the transfer of training to a target in an orthogonal orientation. Strong learning and partial transfer were observed. We tested whether task precision (difficulty) had an effect on template change and transfer: Observers trained in either a high-precision (small, 60° phase difference) or a low-precision task (180°). Task precision did not have an effect on the amount of template change or transfer, suggesting that task precision per se does not determine whether learning generalizes. Classification images show that training made observers use more task-relevant features and unlearn some irrelevant features. The transfer templates resembled partially optimized versions of templates in training sessions. The template change direction resembles ideal learning significantly but not completely. The amount of template change was highly correlated with the amount of learning.

Investigating shape perception by classification images.

Radial frequency (RF) patterns are circular contours where the radius is modulated sinusoidally. These stimuli can represent a wide range of common shapes and have been popular for investigating human shape perception. Theories postulate a multistage model where a global contour integration mechanism integrates the outputs of local curvature-sensitive mechanisms. However, studies on how the local contour features are processed have been mostly based on indirect experimental manipulations. Here, we use a novel way to explore the contour integration, using the classification image (a psychophysical reverse-correlation) method. RF contours were composed of local elements, and random "radial position noise" offsets were added to element radial positions. We analyzed the relationship between trial-to-trial variations in radial noise and corresponding behavioral responses, resulting in a "shape template": an estimate of the contour parts and features that the visual system uses in the shape discrimination task. Integration of contour features in a global template-like model explains our data well, and we show that observer performance for different shapes can be predicted from the classification images. Classification images show that observers used most of the contour parts. Further analysis suggests linear rather than probability summation of contour parts. Convex forms were detected better than concave forms and the corresponding templates had better sampling efficiency. With sufficient presentation time, we found no systematic preferences for a certain class of contour features (such as corners or sides). However, when the presentation time was very short, the visual system might prefer corner features over side features.

Detection of small orientation changes and the precision of visual working memory.

We investigated the precision of orientation representations with two tasks, change detection and recall. Previously change detection has been measured only with relatively large orientation changes compared to psychophysical thresholds. In the first experiment, we measured the observers' ability (d') to detect small changes in orientation (5-30°) with 1-4 Gabor items. With one item even a 10° change was well detected (average d'=2.5). As the amount of change increased to 30°, the d' increased to 5.2. When the number of items was increased, the d's gradually decreased. In the second experiment, we used a recall task and the observers adjusted the orientation of a probe Gabor to match the orientation of a Gabor held in the memory. The standard deviation (s.d.) of errors was calculated from the Gaussian distribution fitted to the data. As the number of items increased from 1 to 6, the s.d. increased from 8.6° to 19.6°. Even with six items, the observers did not make any random adjustments. The results show a square root relation between the d'/s.d. and the number of items. The d' in change detection is directly proportional to the square root of (1/n) and the orientation change. The increase of the s.d. in recall task is inversely proportional to square root of (1/n). The results suggest that limited resources and precision of representations, without additional assumptions, determine the memory performance.

The oceanic state: a conceptual elucidation in terms of modal contact.

Despite its lengthy history in psychoanalysis, the psychological origins, essential features and value of the oceanic state remain open to dispute. This ambiguity has come at a cost to the clarity of theoretical discussions on the topic. In working towards a conceptual elucidation, the author maintains that there are three primary accounts of the oceanic state: the metaphysical one of Romain Rolland, the developmental one of Sigmund Freud, and the cognitive-perceptual one of Anton Ehrenzweig. Based on the notion of modal contact, he argues that the accounts share a general theoretical structure that establishes as the necessary criterion for all oceanic states the loosening of ego boundaries and sufficient contact between differentiated and undifferentiated modalities of the mind. However, within this common structure, the accounts employ dissimilar metapsychologies to promulgate oceanic states of appreciably distinct kinds. To support this view, the author carries out a comparative examination of the modal contacts involved in the primary accounts' oceanic states. To conclude, he reviews the main implications of the notion of modal contact vis-à-vis recent discussion on oceanic phenomena, and puts forward for consideration a pluralist account of the mind that can accommodate the existence of several kinds of oceanic states.

Visual working memory for amplitude-modulated shapes.

We investigated the trade-off between capacity and precision in visual working memory with two different tasks: delayed discrimination and recall. The stimuli were radial frequency patterns that require global pooling of local visual features. The thresholds in delayed amplitude discrimination were measured with a two-interval, forced-choice setup using the Quest procedure. In the recall experiment, the observers' task was to adjust the amplitude of a probe to match the amplitude of a cued item. For one item, the amplitude thresholds were low (0.01-0.05) and the adjustments precise (standard deviations, 0.03-0.05). As the number of items increased from one to six, there was a linear, 6-to-14-fold increase in the thresholds (0.14-0.29) and a 1.5-to-3-fold increase in the standard deviations (0.06-0.11). No sudden or complete breakdown in performance was observed for any subject. The results show a continuous trade-off between memory capacity and precision; six items can be discriminated with the same performance level (75% correct) as one item if the difference between the stimuli is set accordingly. Thus, the stimulus discriminability determines the capacity of visual working memory, and the trade-off between the capacity and precision is linear.

Human working memory for shapes of radial frequency patterns.

We investigated the human working memory for contour shapes. The memory items were radial frequency patterns (Wilkinson, Wilson, & Habak, 1998), which require global pooling of local contour orientation at the 'intermediate' levels of visual system. We used change-detection paradigm and d' measure to determine the forgetting of a single pattern as a function of a retention interval, and the storage capacity for several simultaneously presented patterns. Results showed that the memory trace of a single shape is not so robust as the representations of simple features. Further, the working memory capacity for contour shapes was very low: just one item could be retained accurately.

Visual features underlying perceived brightness as revealed by classification images.

Along with physical luminance, the perceived brightness is known to depend on the spatial structure of the stimulus. Often it is assumed that neural computation of the brightness is based on the analysis of luminance borders of the stimulus. However, this has not been tested directly. We introduce a new variant of the psychophysical reverse-correlation or classification image method to estimate and localize the physical features of the stimuli which correlate with the perceived brightness, using a brightness-matching task. We derive classification images for the illusory Craik-O'Brien-Cornsweet stimulus and a "real" uniform step stimulus. For both stimuli, classification images reveal a positive peak at the stimulus border, along with a negative peak at the background, but are flat at the center of the stimulus, suggesting that brightness is determined solely by the border information. Features in the perceptually completed area in the Craik-O'Brien-Cornsweet do not contribute to its brightness, nor could we see low-frequency boosting, which has been offered as an explanation for the illusion. Tuning of the classification image profiles changes remarkably little with stimulus size. This supports the idea that only certain spatial scales are used for computing the brightness of a surface.

From local to global: Cortical dynamics of contour integration.

Processing of global contours requires integration of local visual information. To study the involvement of different cortical areas and the temporal characteristics of their activity in such integration, we recorded neuromagnetic responses to arrays of Gabor patches in which a proportion of the patches was oriented either tangentially or radially with respect to a global circular contour; arrays with random patch orientations served as control stimuli. The first responses at 60-80 ms around the calcarine sulcus were similar to all stimuli. Starting from 130 ms, responses to the tangential contours differed significantly from responses to control stimuli, and the difference reached its maximum at 275 ms. The most pronounced differences emerged around the parieto-occipital sulcus, precuneus, cuneus, and superior and middle occipital gyri. This pattern of cortical activity was similar irrespective of whether the local elements were radial or tangential to the circle; however, the differences were smaller for the radial contours and tended to start 20-30 ms later. Correspondingly, discrimination reaction times were shortest for the contours consisting of tangential elements. These results demonstrate two spatially and temporally distinct stages of visual cortical processing, the first one limited to local features and the second one integrating information at a more global level.

Detection of irregular spatial structures.

Wilson et al.'s (1997) study on Glass patterns suggested that the integration of stimulus features into a linear shape occurs quite locally, whereas curved structures--such as circular--require global summation. Their conclusion was based on experiments in which they varied the size of the signal area containing a spatial structure. In the present study, we tested the integration of constant-sized linear and curved Glass patterns by varying their global irregularity. If the mechanisms underlying the detection of a Glass pattern pool features globally throughout the stimulus, the irregularity should have a strong effect on detection performance. The irregular Glass patterns were composed of a variable number of sub-areas, each of which contained its own linear or curved structure. The structural irregularity impaired the detection of the curved patterns, whereas the thresholds for the linear patterns were not affected. Thus, our results are in line with the notion that the integration of curved Glass patterns occurs more globally than the integration of linear patterns.

Perception of mirror symmetry in amblyopic vision.

Mirror symmetry is ubiquitous in natural visual scenes, and detection of mirror symmetry seems to be a global, automatic, effortless and important aspect of visual perception. The perception of mirror symmetry has not been studied in humans with amblyopia. In this paper we measured and quantified the detection of mirror symmetry in adults with naturally occurring amblyopia. Our results show that amblyopia may severely impair the detection of mirror symmetry, and that this impairment is not simply a consequence of reduced stimulus visibility. Rather, we suggest that this loss may reflect, at least in part, a deficit in the integration of local orientation information.

Identification of p100 as a coactivator for STAT6 that bridges STAT6 with RNA polymerase II.

STAT6 is a central mediator of IL-4-induced gene responses. STAT6-mediated transcription is depend ent on the C-terminal transcription activation domain (TAD), but the mechanisms by which STAT6 activates transcription are poorly understood. Here, we have identified the staphylococcal nuclease (SN)-like domain and tudor domain containing protein p100 as a STAT6 TAD interacting protein. p100 was originally characterized as a transcriptional coactivator for Epstein-Barr virus nuclear antigen 2. STAT6 interacted with p100 in vitro and in vivo. The interaction was mediated by the TAD domain of STAT6 and the SN-like domain of p100. p100 did not affect the immediate activation events of STAT6, but enhanced STAT6-mediated transcriptional activation and the IL-4-induced Igepsilon gene transcription in human B-cell line. Finally, p100 associated with the large subunit of RNA polymerase II and was mediating interaction between STAT6 and RNA polymerase II. These findings identify p100 as a novel coactivator for STAT6 and suggest that p100 functions as a bridging factor between STAT6 and the basal transcription machinery.