Spatial frequency channels in human vision and the threshold for adaptation

Serial dependence in position occurs at the time of perception

Observers perceive objects in the world as stable over space and time, even though the visual experience of those objects is often discontinuous and distorted due to masking, occlusion, camouflage, or noise. How are we able to easily and quickly achieve stable perception in spite of this constantly changing visual input? It was previously shown that observers experience serial dependence in the perception of features and objects, an effect that extends up to 15 seconds back in time. Here, we asked whether the visual system utilizes an object's prior physical location to inform future position assignments in order to maximize location stability of an object over time. To test this, we presented subjects with small targets at random angular locations relative to central fixation in the peripheral visual field. Subjects reported the perceived location of the target on each trial by adjusting a cursor's position to match its location. Subjects made consistent errors when reporting the perceived position of the target on the current trial, mislocalizing it toward the position of the target in the preceding two trials (Experiment 1). This pull in position perception occurred even when a response was not required on the previous trial (Experiment 2). In addition, we show that serial dependence in perceived position occurs immediately after stimulus presentation, and it is a fast stabilization mechanism that does not require a delay (Experiment 3). This indicates that serial dependence occurs for position representations and facilitates the stable perception of objects in space. Taken together with previous work, our results show that serial dependence occurs at many stages of visual processing, from initial position assignment to object categorization.

Orientation and Spatial Frequency Selectivity following Adaptation: A Reaction Time Study

The aim of the study was to determine orientation and spatial frequency sensitivity using reaction times (RTs) in an adaptation paradigm. Simple RTs were measured to the onset of a Gabor patch (SD = 1.2 deg, spatial frequency = 4 cycles deg−1). Observers adapted for 10 s to a 4 cycles deg−1 grating presented at a series of orientations (0, 2, 5, 10, 22.5, 45, 90°) or spatial frequencies (±0.5, 1, and 2 octaves). The contrast of the test grating was 4x each participant's unadapted threshold. The effect of adaptation was evaluated by transforming RTs to effective contrast reduction using RT-based contrast response functions. RTs increased by between ∼ 100 ms to 150 ms when the test and adapting gratings were of the same orientation or spatial frequency. The effect became less pronounced as the difference in orientation or spatial frequency increased. The average bandwidths for orientation and spatial frequency were 17.4° and 1.24 octaves, respectively. The method has some advantages over traditional approaches. It reveals a rapid time course of adaptation recovery with a half-life of about 13 s to 23 s. RTs form a rapid and easily implemented technique for assessing the underlying physiological mechanisms that control adaptation at suprathreshold levels of contrast.

The temporal course of recovery from brief (sub-second) adaptations to spatial contrast

Visual adaptation is a critical and ubiquitous mechanism that occurs for any stimulus feature and involves a continuous adjustment of the neuronal contrast gain. These adjustments prevent our visual system from dropping in sensitivity for the prevailing ranges of stimulus features that are processed at a given time. In addition to the classical adaptation, which arises over several seconds to minutes, a number of psychophysical, electrophysiological and interference studies have documented a much faster form of adaptation occurring with motion stimuli. This faster adaptation operates on a sub-second scale. In the present study, we investigated whether a fast form of adaptation also exists for spatial contrast and whether its characteristics (e.g., dependence on the duration of adaptation, time course of recovery) are similar to the classical, slower contrast adaptation. We found that a fast form of adaptation does exist and is maximal at intervals of 16-50 ms after the offset of the adapting stimulus. Similar to what previous studies have found regarding the classical contrast adaptation, the initial threshold elevation of this study did not depend on the duration of the adapting stimulus, but only on its contrast. Our results showed that the function which best describes the decay of brief adaptations to high-contrast stimuli was a double exponential decay function, whereas the best function for describing adaptation to low-contrast stimuli was a single exponential decay function with a very fast recovery rate. Thus, adapting contrast influences both the threshold elevation, which rises with increasing adapting contrast, and the time course of recovery from adaptation. Overall, our data suggest the presence of a mechanism that is similar to the classical contrast adaptation involved in longer adaptations, but it operates over much shorter timescales.

Spatial-frequency thresholds for object categorisation at basic and subordinate levels

In an attempt to understand how low-level visual information contributes to object categorisation, previous studies have examined the effects of spatially filtering images on object recognition at different levels of abstraction. Here, the quantitative thresholds for object categorisation at the basic and subordinate levels are determined by using a combination of the method of adjustment and a match-to-sample method. Participants were asked to adjust the cut-off of either a low-pass or high-pass filter applied to a target image until they reached the threshold at which they could match the target image to one of six simultaneously presented category names. This allowed more quantitative analysis of the spatial frequencies necessary for recognition than previous studies. Results indicate that a more central range of low spatial frequencies is necessary for subordinate categorisation than basic, though the difference is small, at about 0.25 octaves. Conversely, there was no effect of categorisation level on high-pass thresholds.

Forward pattern masking and adaptation: effects of duration, interstimulus interval, contrast, and spatial and temporal frequency

Eight experiments are described that compare pattern adaptation and forward pattern masking by examining the effects of five variables on the contrast threshold of a target presented after an adapter or masker. The target is a Gabor pattern with a center frequency of 2 c/deg and a duration of 33 msec. Thresholds are determined using an adaptive spatial forced-choice method. Principal results are as follows. (1) An adapt-refresh regime with a 2 sec refresh and a 2 sec recovery period on each trial is shown to maintain constant performance. (2) Desensitization is very rapid, reaching near maximum in < 200 msec. (3) Recovery is very rapid during the first 100-200 msec and then very slow with the rate of slow recovery decreasing as adapter/masker duration increases. (4) Threshold vs contrast functions are step-like for certain frequency pairs. (5) Sensitivity vs frequency functions derived from adapting and masking are similar in form. (6) Masker temporal frequency (0-15 Hz) has very little effect. These results are described by a theory that postulates that the target is detected by a few mechanisms that are differentially tuned to spatial frequency. The effect of both a forward masker and an adapter is to desensitize the mechanisms that respond to it. Recovery is a weighted sum of two decay processes, one fast and one slow. The theory fits the data from both paradigms well with some differences in parameters.

Measurement of visual channels by contrast adaptation

Inspection of a high-contrast grating pattern affects our ability to detect patterns that are similar. This technique can be used to infer the underlying mechanisms of the visual system. By using this technique, measurements of the bandwidth of orientation channels are taken for different levels of adapting contrast and adapting duration. If the threshold elevation is plotted as the difference between the unadapted and adapted threshold in decibels, then the orientation bandwidth is invariant if taken at some fraction of the maximum elevation. This results from the fact that, as the orientation difference between the adapting and test patterns increases, the function relating threshold elevation to adapting contrast reduces in slope. These data contradict the often-used 'equivalent contrast transformation' (in which the fall off in the adaptation effect with respect to orientation is expressed in terms of an equivalent reduction in adapting contrast) as this would produce quite different bandwidths at different adapting contrasts. The data also address the issue of the neuronal mechanisms of adaptation.

Adaptation in single units in visual cortex: the tuning of aftereffects in the temporal domain

Adaptation-induced changes in the temporal-frequency tuning and direction selectivity of cat visual cortical cells were studied. Aftereffects were induced largely independent of direction. Adapting in either direction reduced responses in both directions. Aftereffects in the direction opposite that adapted were only slightly weaker than were aftereffects in the adapted direction. No cell showed any enhancement of responses to drifting test stimuli after adapting with moving gratings. Adapting in a cell's null direction usually had no effect. Dramatic differences between the adaptation characteristics of moving and stationary stimuli were observed, however. Furthermore, aftereffects were temporal frequency specific. Temporal frequency-specific aftereffects were found in both directions: adapting in one direction induced frequency-specific effects in both directions. This bidirectionality of frequency-specific aftereffects applied to the spatial domain as well. Often, aftereffects in the direction opposite that adapted were more narrowly tuned. In general, adaptation could shift a cell's preferred temporal frequency. Aftereffects were most prominent at high temporal frequencies when testing in the adapted direction. Aftereffects seemed to be more closely linked to temporal frequency than to velocity matching. These results constrain models of cortical connectivity. In particular, we argue against schemes by which direction selectivity is generated by inhibiting a cell specifically when stimulated in the nonpreferred direction. Instead, we argue that cells receive bidirectional spatially and temporally tuned inputs, which could combine in spatiotemporal quadrature to produce direction selectivity.

Orientation and spatial frequency selectivity of adaptation to color and luminance gratings

Prolonged viewing of sinusoidal luminance gratings produces elevated contrast detection thresholds for test gratings that are similar in spatial frequency and orientation to the adaptation stimulus. We have used this technique to investigate orientation and spatial frequency selectivity in the processing of color contrast information. Adaptation to isoluminant red-green gratings produces elevated color contrast thresholds that are selective for grating orientation and spatial frequency. Only small elevations in color contrast thresholds occur after adaptation to luminance gratings, and vice versa. Although the color adaptation effects appear slightly less selective than those for luminance, our results suggest similar spatial processing of color and luminance contrast patterns by early stages of the human visual system.

Dichoptic interaction of harmonically related spatial and temporal frequencies

Steady-state visual evoked potentials were recorded in response to contrast Reprint requests: Dr. Elmar T. Schmeisser, Division of Ocular Hazards, Letterman Army complex (sum of two spatial frequencies) suprathreshold sinusoidal gratings of two and six cycles/degree in a 5 degrees field. Sine mode counterphasing enhances the second temporal harmonic component of the evoked waveform; square mode counterphasing generates a response with essentially all the power at the fundamental temporal frequency. Digital sums of observed responses to simple gratings presented alone had significantly greater amplitudes than any observed response to presented complex gratings. Response amplitudes ranked significantly by condition: binocular greater than dichoptic greater than monoptic. It is concluded that (a) sinusoidal counterphasing stimuli lead to two stimulus events per shift (an observed grating onset and offset), and (b) spatial frequency channels that are harmonically related in 1:3 ratio do not destructively interfere with each other in either monoptic or dichoptic presentation.

An action spectrum for spatial-frequency adaptation

The mechanism of spatial-frequency adaptation, and channel-bandwidth estimates derived from adaptation, were examined in a series of three experiments. The first two experiments measured adaptation as a function of both the frequency and modulation of the adapt stimulus. The threshold for producing adaptation at different adapt frequencies is considered to be an action spectrum whose half-amplitude bandwidth is found to lie between 1/3 and 2/3 octave. The third experiment employed a novel adaptation paradigm--varying the adapt contrast until the test grating was no longer perceptible--but the results supported those of the earlier experiments. It is concluded that adaptation is probably due to a fatigue-like mechanism, and that a sensitivity measure of adaptation (action spectrum) is probably a better representation of underlying channel tuning than is the more customary response measure.

Is spatial adaptation an after-effect of prolonged inhibition?

1. The elevation of the thresholds for sinusoidal gratings of 4 c/deg and 6.7 c/deg was examined after adapting to gratings of 4 c/deg. Threshold elevation was determined as a function of the adapting contrast.2. It was confirmed that, when the adapting and testing spatial-frequencies were the same, the threshold elevation curve extrapolated to zero elevation at an adapting contrast equal to the pre-adaptation threshold for the adapting grating.3. It was found that the threshold elevation curve for 6.7 c/deg gratings also extrapolated to zero elevation at the pre-adaptation threshold for the adapting grating, even though the testing and adapting frequencies were different.4. It is argued that the latter result shows that adaptation is not simply an after-effect of prolonged excitation of a channel. Threshold elevation curves may represent the bandwidths of inhibitory interactions between channels.