Learning in grating waveform discrimination: Specificity for orientation and spatial frequency
Abstract
The effects of practice in the discrimination of briefly flashed gratings were investigated by a forced-choice procedure with error correction in a number of tasks requiring discimination either of pairs of complex gratings of different waveforms or of “simple” (sinusoidal) gratings of slightly different spatial frequency. The percentage of correct responses progressively increases with repetition of trials up to 100–200 trials and then levels off, remaining rather constant thereafter even after days or weeks, in all tasks involving discrimination of complex gratings.
However, when the gratings are set perpendicular to those used for the training sessions, or their spatial frequency is changed by 1 octave, the effects of previous perceptual learning are lost, while transfer of learning effects is obtained for smaller changes in orientation (±30°) or spatial frequency (±12 octave). The spatial frequency discrimination of sinusoidal gratings does not improve with a comparable number of trials.
Reference (8)
- BraddickO. et al.
Channels in vision: basic aspects
- CavanaghP.
Size and position invariance in the visual system
Perception
(1977) - CampbellF.W. et al.
Spatialfrequency discrimination in human vision
J. opt. Soc. Am.
(1970) - FiorentiniA. et al.
Perceptual learning specific for orientation and spatial frequency
Nature
(1980)
Cited by (232)
Profiles of visual perceptual learning in feature space
2024, iScienceVisual perceptual learning (VPL), experience-induced gains in discriminating visual features, has been studied extensively and intensively for many years, its profile in feature space, however, remains unclear. Here, human subjects were trained to perform either a simple low-level feature (grating orientation) or a complex high-level object (face view) discrimination task over a long-time course. During, immediately after, and one month after training, all results showed that in feature space VPL in grating orientation discrimination was a center-surround profile; VPL in face view discrimination, however, was a monotonic gradient profile. Importantly, these two profiles can be emerged by a deep convolutional neural network with a modified AlexNet consisted of 7 and 12 layers, respectively. Altogether, our study reveals for the first time a feature hierarchy-dependent profile of VPL in feature space, placing a necessary constraint on our understanding of the neural computation of VPL.
A distinct route for efficient learning and generalization in autism
2022, Current BiologyVisual skill learning is the process of improving responses to surrounding visual stimuli.1 For individuals with autism spectrum disorders (ASDs), efficient skill learning may be especially valuable due to potential difficulties with sensory processing2 and challenges in adjusting flexibly to changing environments.3,4 Standard skill learning protocols require extensive practice with multiple stimulus repetitions,5, 6, 7 which may be difficult for individuals with ASD and create abnormally specific learning with poor ability to generalize.4 Motivated by findings indicating that brief memory reactivations can facilitate skill learning,8,9 we hypothesized that reactivation learning with few stimulus repetitions will enable efficient learning in individuals with ASD, similar to their learning with standard extensive practice protocols used in previous studies.4,10,11 We further hypothesized that in contrast to experience-dependent plasticity often resulting in specificity, reactivation-induced learning would enable generalization patterns in ASD. To test our hypotheses, high-functioning adults with ASD underwent brief reactivations of an encoded visual learning task, consisting of only 5 trials each instead of hundreds. Remarkably, individuals with ASD improved their visual discrimination ability in the task substantially, demonstrating successful learning. Furthermore, individuals with ASD generalized learning to an untrained visual location, indicating a unique benefit of reactivation learning mechanisms for ASD individuals. Finally, an additional experiment showed that without memory reactivations ASD subjects did not demonstrate efficient learning and generalization patterns. Taken together, the results provide proof-of-concept evidence supporting a distinct route for efficient visual learning and generalization in ASD, which may be beneficial for skill learning in other sensory and motor domains.
Grating orientation task trial numbers for short- and long-term tactile discrimination learning
2021, Journal of Clinical NeurosciencePerceptual learning generally improves with training intensity, but the number of training trials sufficient for transient and long-term improvement in tactile grating orientation task (GOT) discrimination has not been systematically studied. To define reliable trial numbers for tactile discrimination learning, we compared tactile orientation discrimination performance of the right-finger following 200 and 400 training trials. Fifty-one neurologically healthy subjects were recruited. Tactile spatial acuity for orientation (parallel or orthogonal to the long axis of the finger) across different grating frequencies was assessed before, immediately after, 30 min after, and 24 h after sessions consisting of 200 training trials (50/block × 4 blocks), 400 training trials (50/block × 8 blocks), or no training (sham control). Both the 200- and 400-trial training conditions reduced the grating orientation discrimination threshold at 24 h after training. In contrast, the control condition had no effect on the grating orientation discrimination threshold. There was a negative correlation between the baseline grating orientation discrimination threshold and training-induced change in threshold (improvement) following both 200 and 400 trials. Fewer GOT trials (200) substantially prolong tactile discrimination learning, presumably by promoting the consolidation of the underlying neuroplastic mechanisms. In this widely used perceptual learning paradigm, 200 and 400 training trials appear effective for inducing short-term and long-term perceptual memory.
Bottom-up and top-down factors of motion direction learning transfer
2019, Consciousness and CognitionPerceptual learning of motion discrimination has long been believed to be motion direction specific. However, recent studies using a double-training paradigm, in which the to-be-transferred condition was experienced through practicing an irrelevant task, found that perceptual learning in various visual tasks, including motion direction discrimination, can transfer completely to new conditions. This transfer occurred when the transfer stimulus was subconsciously presented, or when top-down attention was allocated to the transfer stimulus (which was absent). In the current study, observers were exposed subconsciously, or directed top-down attention, to the transfer motion direction, either simultaneously or successively with training. Data showed that motion direction learning transferred to the transfer direction, and suggest that motion direction learning specificity may result from under-activations of untrained visual neurons due to insufficient bottom-up stimulation and/or lack of top-down attention during training. These results shed new light on the neural mechanisms underlying motion perceptual learning and provide a constraint for models of motion perceptual learning.
The effect of training with inverted faces on the selective use of horizontal structure
2019, Vision ResearchA growing body of evidence demonstrates that selective processing of structure conveyed by horizontally oriented spatial frequency components is associated with upright face discrimination accuracy and the magnitude of the face inversion effect. In this study, we examined whether the increase in discrimination accuracy for inverted faces that is known to result from practice would coincide with more selective processing of horizontal structure in inverted faces. To assess this hypothesis, our observers practiced discrimination of inverted faces for three training sessions and we measured accuracy, efficiency relative to an ideal observer, and horizontal selectivity before and after training. As hypothesized, we observed more efficient discrimination and more selective processing of horizontal structure after training. However, the effects of training did not generalize reliably to novel face exemplars.
Perceptual learning while preparing saccades
2018, Vision ResearchTraditional perceptual learning protocols rely almost exclusively on long periods of uninterrupted fixation. Taking a first step towards understanding perceptual learning in natural vision, we had observers report the orientation of a briefly flashed stimulus (clockwise or counterclockwise from a reference orientation) presented strictly during saccade preparation at a location offset from the saccade target. For each observer, the saccade direction, stimulus location, and orientation remained the same throughout training. Subsequently, we assessed performance during fixation in three transfer sessions, either at the trained or at an untrained location, and either using an untrained (Experiment 1) or the trained (Experiment 2) stimulus orientation. We modeled the evolution of contrast thresholds (i.e., the stimulus contrast necessary to discriminate its orientation correctly 75% of the time) as an exponential learning curve, and quantified departures from this curve in transfer sessions using two new, complementary measures of transfer costs (i.e., performance decrements after the transition into the Transfer phase). We observed robust perceptual learning and associated transfer costs for untrained locations and orientations. We also assessed if spatial transfer costs were reduced for the remapped location of the pre-saccadic stimulus—the location the stimulus would have had (but never had) after the saccade. Although the pattern of results at that location differed somewhat from that at the control location, we found no clear evidence for perceptual learning at remapped locations. Using novel, model-based ways to assess learning and transfer costs, our results show that location and feature specificity, hallmarks of perceptual learning, subsist if the target stimulus is presented strictly during saccade preparation throughout training.