Jeounghoon Kim

A model for motion coherence and transparency

A recent model for two-dimensional motion processing in MT has demonstrated that perceived direction can be accurately predicted by combining Fourier and non-Fourier component motion signals using a vector sum computation. The vector sum direction is computed by a neural network that weights Fourier and non-Fourier components by the cosine of the component direction relative to that of each pattern unit, after which competitive inhibition extracts the signals of the most active units. It is shown here that a minor modification of the connectivity in this network suffices to predict transitions from motion coherence to transparency under a wide range of circumstances. It is only necessary that the cosine weighting function and competitive inhibition be limited to directions within +/- 120 deg of each pattern units preferred direction. This network responds by signaling one pattern direction for coherent motion but two distinct directions for transparent motion. Based on this, neural networks with properties of MT and MST neurons can automatically signal motion coherence or transparency. In addition, the model accurately predicts motion repulsion under transparency conditions.

Dependence of plaid motion coherence on component grating directions

We measured motion coherence for plaids composed of two different spatial frequency (SF) cosine grating components moving at various relative angles. The component SFs were in a ratio of 6:1, and several component motion directions were chosen to produce small to large angular differences. For angles of less than +/- 45 deg all subjects perceived rigid coherent motion, while for angles of +/- 45 deg or greater the components were perceived to slide transparently. The results were not altered when we changed the ratio of component contrasts or speeds over a factor of 2 and varied the SF ratio up to 9:1. These results show that transparency or coherence of moving plaids in different spatial scales depends on the relative component motion directions and is relatively independent of contrast, speed, and SF difference between the components. This angular dependence also explains recent data previously thought to be based on a visual computation of multiplicative transparency. A quantitative model in which the resultant motion on each scale provides a facilitative biasing signal to units tuned to similar directions (within +/- 30 deg) on other scales explains the experimental results.

Perceived motion in the vector sum direction

The visual system is known to extract both Fourier (e.g. contour) and non-Fourier (e.g. texture boundary) motion signals from moving patterns. Recent evidence suggests that both Fourier and non-Fourier motion signals are involved in the analysis of rigid two-dimensional image motion. To determine how these signals are combined, we have measured perceived directions for plaids composed of either two non-Fourier components or one Fourier plus one non-Fourier component. In all cases the visual system combines these components to produce perceived motion in the vector sum direction, even when this deviates by as much as 53 deg from the intersection of constraints direction. This result was predicted by a model for pattern motion that combines Fourier and non-Fourier component motion signals using a vector sum operation.

Direction repulsion between components in motion transparency

We measured the perceived direction of one motion component as a function of the contrast and speed of a second component for three pattern classes: plaids with two different spatial frequency components, multi-aperture patterns, and contrast-modulated (CM) patterns. The components were moving at +/- 63.4 or +/- 71.6 deg to the vertical, angles where motion transparency always occurred under our conditions. For multi-aperture and CM patterns on a single spatial scale, the components were perceived to deviate from the component motion directions by up to 20 deg at high contrasts or high speeds of the second component. However, for plaids with components on different spatial scales, the test components were perceived moving in the component directions regardless of the contrast or the speed of the second component. Our data show that this direction repulsion between components occurs within a single spatial scale but not between widely separated spatial scales. This implies that two different mechanisms are involved in motion transparency.

Motion Integration over Space: Interaction of the Center and Surround Motion

Motion integration occurs over a restricted range of visual space. However, there have been studies suggesting interactions among motion detectors operating on widely separated spatial regions. To understand these lateral spatial interactions beyond motion pooling regions, we examined the effect of surrounding motion on the direction of the center stimulus under several stimulus conditions. We have found that there is a motion direction shift of the center stimulus caused by surrounding motion depending on its motion direction, spatial proximity to the center stimulus, contrast, speed, and the extent of motion area. This effect was observed both for monocular and dichoptic presentations of the pattern. However, the perceived direction shift decreased when the spatial frequency ratio of the center and surround stimuli varied, or a non-Fourier motion pattern was used for both center and surround stimuli. We present a model consisting of lateral inhibitory interactions between pattern motion unit networks to explain the direction shift observed in the experiments.

A Functional Dissociation of Conflict Processing Within Anterior Cingulate Cortex

Goal‐directed behavior requires cognitive control to regulate the occurrence of conflict. The dorsal anterior cingulate cortex (dACC) has been suggested in detecting response conflict during various conflict tasks. Recent findings, however, have indicated not only that two distinct subregions of dACC are involved in conflict processing but also that the conflict occurs at both perceptual and response levels. In this study, we sought to examine whether perceptual and response conflicts are functionally dissociated in dACC. Thirteen healthy subjects performed a version of the Stroop task during functional magnetic resonance imaging (fMRI) scanning. We identified a functional dissociation of the caudal dACC (cdACC) and the rostral dACC (rdACC) in their responses to different sources of conflict. The cdACC was selectively engaged in perceptual conflict whereas the rdACC was more active in response conflict. Further, the dorsolateral prefrontal cortex (DLPFC) was coactivated not with cdACC but with rdACC. We suggest that cdACC plays an important role in regulative processing of perceptual conflict whereas rdACC is involved in detecting response conflict. Hum Brain Mapp, 2011.

Multiple cognitive control mechanisms associated with the nature of conflict

Cognitive control is required to regulate conflict. The conflict monitoring theory suggests that the dorsal anterior cingulate cortex (dACC) is involved in detecting response conflict and the dorsolateral prefrontal cortex (DLPFC) plays a critical role in regulating conflict. Recent studies, however, have suggested that rostral dACC (rdACC) responds to response conflict whereas caudal dACC (cdACC) is associated with perceptual conflict. Moreover, DLPFC has been engaged only in regulation of response conflict. A neural network involved in perceptual conflict, however, remains unclear. In this study, we used functional magnetic resonance imaging (fMRI) in an attempt to reveal monitor-controller networks corresponding to either perceptual conflict or response conflict. A version of the Stroop color matching task was used to manipulate perceptual conflict, response conflict was manipulated by an arrow. The results demonstrated that rdACC and DLPFC were engaged in response conflict whereas cdACC and the dorsal portion of premotor cortex (pre-PMd) were involved in perceptual conflict. Interestingly, the posterior parietal cortex (PPC) was activated by both types of conflict. Correlation analyses between behavioral conflict effects and neural responses demonstrated that rdACC and DLPFC were associated with response conflict whereas cdACC and pre-PMd were associated with perceptual conflict. PPC was not correlated with either perceptual conflict or response conflict. We suggest that cdACC and pre-PMd play critical roles in perceptual conflict processing, and this network is independent from the rdACC/DLPFC network for response conflict processing. We also discussed the function of PPC in conflict processing.

Dynamics of a divisive gain control in human vision.

Evidence for a divisive contrast gain control in human vision was obtained using a contrast version of the probe-on-flash technique that has been employed in the light adaptation literature. Thresholds were measured for a briefly flashed (30 ms), vertical test pattern superimposed on a cosine mask as a function of time after mask onset (SOA). Threshold elevations declined monotonically for SOAs up to 150 ms. and exhibited an exponential time course with an average time constant of 51 ms. Increment thresholds for the test as a function of mask contrast provide direct evidence that these effects are due to operation of a divisive gain control within the first 150 ms after stimulus onset. Experiments to measure the spatial spread of this gain control show it to be localized to a region of no more than 45 arc min radius.

Conflict adjustment through domain-specific multiple cognitive control mechanisms

Cognitive control is required to regulate conflict between relevant and irrelevant information. Although previous neuroimaging studies have focused on response conflict, recent studies suggested that distinct neural networks are recruited in regulating perceptual conflict. The aim of the current study was to distinguish between brain areas involved in detecting and regulating perceptual conflict using a conflict adjustment paradigm. The Stroop color-matching task was combined with an arrow version of the Stroop task in order to independently manipulate perceptual and response conflicts. Behavioral results showed that post-conflict adjustment for perceptual and response conflicts were independent from each other. Imaging results demonstrated that the caudal portion of the dorsal cingulate cortex (cdACC) was selectively associated with the occurrence of perceptual conflict, whereas the left dorsal portion of the premotor cortex (pre-PMd) was selectively associated with both preceding and current perceptual conflict trials. Furthermore, the rostral portion of the dorsal cingulate cortex (rdACC) was selectively linked with response conflict, whereas the left dorsolateral prefrontal cortex (DLPFC) was selectively involved in both preceding and current response conflict trials. We suggest that cdACC is involved in detecting perceptual conflict and left pre-PMd is involved in regulating perceptual conflict, which is analogous to the recruitment of rdACC and left DLPFC in control processes for response conflict. Our findings provide support for the hypothesis that multiple independent monitor-controller loops are implemented in the frontal cognitive control system.

Task-dependent response conflict monitoring and cognitive control in anterior cingulate and dorsolateral prefrontal cortices

Previous experience affects our behavior in terms of adjustments. It has been suggested that the conflict monitor-controller system implemented in the prefrontal cortex plays a critical role in such adjustments. Previous studies suggested that there exists multiple conflict monitor-controller systems associated with the level of information (i.e., stimulus and response levels). In this study, we sought to test whether different types of conflicts occur at the same information processing level (i.e., response level) are independently processed. For this purpose, we designed a task paradigm to measure two different types of response conflicts using color-based and location-based conflict stimuli and measured the conflict adaptation effects associated with the two types of conflicts either independently (i.e., single conflict conditions) or simultaneously (i.e., a double-conflict condition). The behavioral results demonstrated that performance on current incongruent trials was faster only when the preceding trial was the same type of response conflict regardless of whether they included a single- or double-conflict. Imaging data also showed that anterior cingulate and dorsolateral prefrontal cortices operate in a task-specific manner. These findings suggest that there may be multiple monitor-controller loops for color-based and location-based conflicts even at the same response level. Importantly, our results suggest that double-conflict processing is qualitatively different from single-conflict processing although double-conflict shares the same sources of conflict with two single-conflict conditions.