Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2013 May 15.
Published in final edited form as: Vision Res. 2011 Jul 26;61:33–38. doi: 10.1016/j.visres.2011.07.019

Task relevancy and demand modulate double-training enabled transfer of perceptual learning

Rui Wang 1, Jun-Yun Zhang 1, Stanley A Klein 2, Dennis M Levi 2, Cong Yu 1
PMCID: PMC3227777  NIHMSID: NIHMS314661  PMID: 21820004

Abstract

Location-specific perceptual learning can be rendered transferrable to a new location with double training, in which feature training (e.g., contrast) is accompanied by additional location training at the new location even with an irrelevant task (e.g. orientation). Here we investigated the impact of relevancy (to feature training) and demand of location training tasks on double training enabled learning transfer. We found that location training with an irrelevant task (Gabor vs. letter judgment, or contrast discrimination) limited transfer of Vernier learning to the trained orientation only. However, performing a relevant suprathreshold orthogonal Vernier task prompted additional transfer to an untrained orthogonal orientation. In addition, the amount of learning transfer may depend on the demand of location training as well as the double training procedure. These results characterize how double training potentiates the functional connections between a learned high-level decision unit and visual inputs from an untrained location to enable transfer of learning across retinal locations.

Keywords: Perceptual learning, double training, location specificity, transfer

1. Introduction

Visual perceptual learning improves discrimination of many basic visual features, such as contrast, orientation, Vernier, and texture. A key feature of it is that learning is often specific to the trained retinal location and orientation (Ahissar & Hochstein, 1997; Crist, et al., 1997; Fahle, 1994; Fahle, 1997; Karni & Sagi, 1991; Saarinen & Levi, 1995; Schoups, et al., 1995; Shiu & Pashler, 1992; Yu, et al., 2004). The location and orientation specificities place important constraints on various perceptual learning models and theories (Adini, et al., 2002; Bejjanki, et al., 2011; Dosher & Lu, 1998; Law & Gold, 2009; Mollon & Danilova, 1996; Poggio, et al., 1992; Teich & Qian, 2003; Zhaoping, et al., 2003), as it is suggested that modeling the neural mechanisms underlying perceptual learning must account for these specificities (Tsodyks & Gilbert, 2004).

However, in recent studies we demonstrated that location specificity and orientation specificity can be decoupled from perceptual learning in a variety of visual tasks with appropriate training procedures (Xiao, et al., 2008; Zhang, et al., 2010a, b). For example, with a feature-plus-location double training procedure, perceptual learning of contrast discrimination (feature training), which is otherwise location specific, can transfer completely to a new location following additional location training at the new location using an irrelevant orientation discrimination task (Xiao et al., 2008). Similarly, with a training-plus-exposure (TPE) procedure, perceptual learning of orientation in foveal vision, which is otherwise specific to the trained orientation, can transfer completely to an orthogonal orientation if an observer is also passively exposed to the orthogonal orientation while performing an irrelevant contrast discrimination task (Zhang et al., 2010a).

These transfer results suggest that at least in some cases perceptual learning is more a general learning process and most likely occurs at a high decision level of information processing beyond the retinotopic and orientation selective visual cortex (Xiao, et al., 2008; Zhang, et al., 2010a; Zhang, et al., 2010b). This argument is supported by a recent fMRI study indicating that the site of perceptual learning may be located in the human medial front cortex, which is also the site of perceptual decision making (Kahnt, et al., 2011). We thus proposed a rule-based learning theory to explain visual perceptual learning and its specificity and transfer (Zhang et al., 2010a). This theory posits that a high-level decision unit learns the rules for performing a visual task through training. However, the learned rules cannot be applied to a new location or orientation automatically because the decision unit cannot functionally connect to the visual inputs representing the new location or orientation with sufficient strength. These inputs are unattended or even suppressed during training when attention is allocated to the trained location or orientation (Gal, et al., 2009; Sylvester, et al., 2009). It is double training and TPE training that reactivate these new inputs, so that the functional connections can be strengthened to enable rule application and transfer of learning.

The current study manipulated the task relevancy (to feature training) and the demand of location training and investigates their impact on the transfer of feature (Vernier) learning from a diagonal quadrant of the visual field. We used three levels of task relevance (passive stimulus exposure, irrelevant, and relevant) and three demand levels (passive stimulus exposure, suprathreshold, and near-threshold) with the location training. Here passive stimulus exposure was least task relevant and least demanding, and was therefore used in both categories as baselines. Vernier learning is found to be strictly specific to the trained location (Xiao et al., 2008), so its transfer to a new location would serve as an excellent indicator of the impact of location training under various task relevancy and demand conditions. Our results show that an observer’s active participation in a visual task at the new location, even a very simple one, rather than passive exposure to the stimuli is necessary for Vernier learning to transfer. Moreover, task-irrelevant location training limits learning transfer to only the trained Vernier orientation, but a relevant suprathreshold orthogonal Vernier task allows additional learning transfer to the orthogonal orientation. In addition, the amount of learning transfer may depend on the demand level in the location training task as well as the particular double training procedure (simultaneous or sequential). In the context of our rule-based perceptual learning theory, these results characterize double training potentiating the functional connections between a learned high-level decision unit and new visual inputs from an untrained retinal location, which makes the transfer of perceptual learning possible.

2. Methods

2.1 Observers and apparatus

Forty observers (undergraduate students in their early twenties at Beijing Normal University) with normal or corrected-to-normal vision participated in this study. All were new to psychophysical experiments and were unaware of the purpose of the study. Informed consent was obtained from each observer before data collection.

The stimuli were generated by a Matlab-based WinVis program (Neurometrics Institute, Oakland, CA) and presented on a 21-in. Sony G520 color monitor (2048 pixel × 1536 pixel, 0.19 mm × 0.19 mm per pixel, 75 Hz frame rate). The mean luminance was 50 cd/m2. Luminance of the monitor was linearized by an 8-bit look-up table. Viewing was monocular with one eye covered with a translucent plastic pad, and a chin-and-head rest helped stabilize the head of the observer. The viewing distance was 1.5 m. Experiments were run in a dimly lit room.

2.2 Stimuli

The Vernier stimulus consisted of a pair of identical Gabor patches (Gaussian windowed sinusoidal gratings) on a mean luminance screen background, which was centered in a visual quadrant at 5° retinal eccentricity (Fig. 1A). The two Gabors had the same spatial frequency (3 cpd), fixed phase (90°), standard deviation (2λ), contrast (0.47), orientation (either vertical or horizontal), and a center-to-center distance of 4λ. To form a specific Vernier offset, the position of each Gabor shifted half the Vernier offset away in opposite directions perpendicular to the Gabor orientation.

Figure 1.

Figure 1

Open in a new tab

Transfer of Vernier learning to a diagonal quadrant where an orthogonal Vernier was passively exposed. A. Stimuli. Observers practiced Vernier discrimination at ori1_loc1 while passively exposed to an orthogonal Vernier flashed simultaneously in the diagonal quadrant (ori2_loc2). B. The mean session-by-session threshold changes in the trained condition and pre- and post-training thresholds at the transfer conditions. C. The MPIs of Vernier performance in the trained condition (left bar) and untrained transfer conditions (right three bars).

The same Vernier stimulus was also used for contrast discrimination training with the Vernier offset jittered at ±15 arcmin. The contrasts of the two Gabors were set at 0.47 and 0.47 + ΔC.

2.3 Procedure

Vernier and contrast discrimination thresholds were measured with a single-interval discrimination staircase procedure. In each trial, the stimulus was presented for 200 ms. For Vernier discrimination, an observer judged whether the right Gabor was higher or lower than the left Gabor for a horizontal Vernier, or the lower Gabor was to the left or right of the upper Gabor for a vertical Vernier. For contrast discrimination, an observer judged which of the two Gabors had a higher contrast. A small fixation cross preceded each trial by 400 ms and stayed through the trial. Auditory feedback was given on incorrect responses. Thresholds were estimated using a classical 3-down-1-up staircase rule that resulted in a 79.4% convergence level. Each staircase consisted of four preliminary reversals and six experimental reversals (approximately 50 trials). The step size of the staircase was 0.05 log units. The geometric mean of the experimental reversals was taken as the threshold for each staircase run.

2.4 Statistics

The amount of perceptual learning was quantified as the Percent Improvement (PI), where PI = (ThreshPre − ThreshPost)/ ThreshPre. We hypothesized that thresholds were lower after training in perceptual learning experiments, so that one-tailed hypothesis tests were used.

3. Results

3.1 Baseline: Feature training plus passive exposure to stimuli at the transfer location

We first studied whether passive exposure to stimuli at the transfer location would enable the transfer of Vernier learning. Passive stimulus exposure was least task-relevant and required the least effort by the observer, so this measure provided baselines for the impacts of task relevancy and demand of location training on learning transfer. Sixteen observers practiced Vernier discrimination at a horizontal or vertical orientation in one of four visual quadrants (ori1_loc1) at 5° retinal eccentricity while an orthogonal Vernier in the diagonal quadrant (ori2_loc2) was flashed simultaneously for the same 200 ms duration (Fig. 1A). The offset of the flashed Vernier was randomly set at ±15 arcmin, or approximately ±2.5 times the mean pre-training Vernier threshold. Because nearly all attention was allocated to Vernier discrimination at the trained location, the flashed Vernier was only passively viewed by an observer. Significant learning was evident at the trained orientation and location (ori1_loc1, Mean Percent Improvement (MPI) = 27.3 ± 2.9%, p < 0.001, paired t-test) after six to seven two-hour daily sessions (Fig. 1B & C). However, learning transferred little to the same or orthogonal orientations in the diagonal visual quadrant with passive stimulus exposure (ori1_loc2, MPI = −0.5 ± 4.4%, p = 0.55; ori2_loc2, MPI = 1.2 ± 4.0%, p = 0.39), as well as the orthogonal orientation at the trained location (ori2_loc1, MPI = 1.5 ± 5.8%, p = 0.40) (Fig. 1B & C). These results indicate that mere passive exposure to the stimuli at the transfer location cannot replace active location training to override the specificity of Vernier learning revealed in our previous study with identical stimuli (Xiao et al., 2008).

3.2. Double training: Feature training plus location training with an irrelevant suprathreshold task

Among the sixteen observers in the passive stimulus exposure experiment (Fig. 1), one did not learn (Vernier performance improved by <3%) and two showed significant transfer (the transfer index (TI), defined as the ratio of transfer/training performance improvements, was >0.5). The remaining thirteen showed little or no improvement in performance in the transfer conditions (mean TI = −0.18 +/− 0.14). We were able to call eleven of these thirteen observers back and split them into two groups, each performing a suprathreshold task either irrelevant or relevant to feature learning. The first group of five observers in the current experiment judged whether a pair of Gabors (Fig. 2A, same as the flashing Gabors in Fig. 1A, presented on 80% of the trials,) or an uppercase letter E (20% of the trials) appeared in the diagonal transfer quadrant for four sessions. Here the observers were forced to perform an irrelevant, non-demanding suprathreshold task (mean accuracy = 99.7%) at the transfer location. For these five observers, the previous Vernier training produced significant improvement at ori1_loc1 (MPI = 30.3 ± 2.9%, p < 0.001, Fig. 2B & C), but learning did not transfer to untrained ori1_loc2 (MPI = −9.7 ± 8.0%, p = 0.85), ori2_loc2 (MPI = −1.9 ± 8.1%, p = 0.59), and ori2_loc1 at the trained location (MPI = −9.5 ± 9.4%, p = 0.81) (Fig. 2B & C). After the new Gabors vs. E judgments, Vernier performance was improved at ori1_loc2 (MPI = 14.5 ± 4.0%, p = 0.012), but not at orthogonal ori2_loc2 (MPI = −2.8 ± 4.7%, p = 0.71) and ori2_loc1 (MPI = −9.6 ± 10.4%, p = 0.79) (Fig. 2B & C), indicating that location training with an irrelevant suprathreshold task can enable the transfer of Vernier learning across retinal locations, but the transfer is limited to the trained orientation. On the other hand, the lower MPI at ori1_loc2 (14.5% vs. 30.3%) appeared to indicate partial transfer of learning (p = 0.047 for MPIs at ori1_loc1 vs. ori1_loc2). However, it was unclear whether the partial transfer truly resulted from location training with an irrelevant suprathreshold task, or from a performance ceiling effect at the transfer location (the post-training thresholds were 4.34 ± 0.47 arcmin at ori1_loc1 and 4.50 ± 0.16 arcmin at ori1_loc2, p = 0.37) combined with the lower pre-training thresholds at ori1_loc2 (Fig. 2B). In the latter case the transfer of learning could actually be complete.

Figure 2.

Figure 2

Open in a new tab

Transfer of Vernier learning to a diagonal quadrant trained with an irrelevant suprathreshold task. A. Stimuli. Observers judged whether the stimuli were Gabors or a letter E. B. Sequential procedure: Vernier performance in various training and transfer conditions before and after initial training at ori1_loc1 (replotted from Fig. 1) and after the Gabors-or-E judgments at loc2. C. The MPIs in the trained condition and untrained transfer conditions after two experimental phases. D. Simultaneous procedure: Vernier training at ori1_loc1 and Gabors-or-E judgments at loc2 in alternating blocks in the same session. E. The MPIs in the trained condition (left bar) and untrained transfer conditions (right three bars) after the simultaneous procedure.

In our previous study (Xiao et al., 2008) double training enabled complete transfer of learning regardless of whether feature and location training (with an irrelevant but demanding near-threshold task) was performed simultaneously or sequentially. However, the above results would suggest that demanding location training might be unnecessary if the transfer of learning is indeed complete. Would the above findings of orientation specificity in learning transfer and potentially complete learning transfer be replicable in a simultaneous double training procedure? We had eight new observers practice Vernier discrimination at ori1_loc1 and judge the same Gabors or E in the diagonal quadrant in alternating blocks of trials in a same session for five sessions. Training improved Vernier performance at ori1_loc1 by 29.6 ± 4.0% (p < 0.001), nearly identical to the 30.3 ± 2.9% MPI in Fig. 2b. Vernier learning again transferred to the same trained orientation ori1_loc2 only (MPI = 18.2 ± 6.6%, p = 0.015), but not to orthogonal ori2_loc2 (MPI = 5.1 ± 4.9%, p = 0.17) and ori2_loc1 (MPI = 4.3 ± 5.1%, p = 0.21) (Fig. 2D & E). These results confirmed that the transfer of learning was specific to the trained orientation under the current double training conditions. However, the transfer was clearly partial with lower MPI at ori1_loc2 than at ori1_loc2 (p = 0.014) and similar pre-training thresholds.

3.3. Double training: Feature training plus location training with a relevant suprathreshold task

The second group of six observers who did not show much transfer after passive stimulus exposure (Fig. 1) judged an orthogonal Vernier with the offset set at five times the threshold after passive stimulus exposure at the transfer location for four sessions (Fig. 3A). Since the offset was well above threshold, the task was not demanding (mean accuracy = 99.1%), but this time the task was relevant to feature (Vernier) learning. For these six observers, previous Vernier training produced significant improvement at ori1_loc1 (MPI = 24.1 ± 3.8%, p < 0.001), and learning did not transfer to ori1_loc2 (MPI = 0.3 ± 3.1%, p = 0.47), ori2_loc2 (MPI = 2.4 ± 2.3%, p = 0.18), and ori2_loc1 (MPI = −5.6 ± 8.7%, p = 0.73) (Fig. 3B & C). However, after the suprathreshold Vernier judgments at ori2_loc2, Vernier learning transferred not only to ori1_loc2 (MPI = 15.5 ± 3.1%, p = 0.002), but also to the orthogonal ori2_loc2 (MPI = 20.4 ± 4.6%, p = 0.003) and ori2_loc1 (MPI = 13.2 ± 5.2%, p = 0.026) (Fig. 3B & C). Moreover, after this two-phase double training there was no significant difference in performance improvement among the trained condition (MPI = 18.6 ± 5.0%, p = 0.007) and three transfer conditions (p = 0.30, repeated measures ANOVA).

Figure 3.

Figure 3

Open in a new tab

Transfer of Vernier learning to a diagonal quadrant trained with a relevant suprathreshold Vernier task. A. Stimuli. The suprathreshold Vernier at ori2_loc2 was five times the threshold. B. Sequential procedure: Vernier performance before and after initial training at ori1_loc1, and after suprathreshold Vernier judgments at ori2_loc2. C. The MPIs in the trained condition and untrained transfer conditions after two experimental phases. D. Simultaneous procedure: Vernier discrimination at ori1_loc1 and suprathreshold Vernier judgments at ori2_loc2 in alternating blocks in the same session. E. The MPIs in the trained condition (left bar) and untrained transfer conditions (right three bars) after the simultaneous procedure. F. Control: Pre- and post-training thresholds in three transfer conditions while only the suprathreshold Vernier task was trained.

We again had seven new observers repeat the experiment in a simultaneous procedure. They practiced Vernier discrimination at ori1_loc1 and judged suprathreshold orthogonal Vernier (five times the mean pre-training threshold) at ori2_loc2 in alternating blocks of trials in the same session for five sessions. Consistent with the sequential double training data, Vernier learning at ori1_loc1 (MPI = 38.9 ± 4.3%, p < 0.001, which is higher than the magnitude in the sequential condition described above), transferred to not only ori1_loc2 (MPI = 17.8 ± 5.8%, p = 0.011), but also orthogonal ori2_loc2 (MPI = 20.9 ± 4.9%, p = 0.003) and ori2_loc1 (MPI = 18.7 ± 8.4%, p = 0.034) (Fig. 3D & E). However, as in Fig. 2d, the simultaneous procedure only produced partial transfer (p = 0.015 for the contrast between the training MPI and three transfer MPIs, repeated measures ANOVA). Taking Fig.2 and Fig. 3 results together, there appears to be a trend that location training with a suprathreshold task may enable complete transfer in a sequential double-training procedure, but only partial transfer in a simultaneous procedure.

In addition, we ran a control experiment in which five new observers performed only the same suprathreshold Vernier task (five times the mean pre-training threshold) at ori2_loc2 for four sessions. This suprathreshold task had no impact on the Vernier thresholds at ori1_loc2, ori2_loc2, and ori2_loc1 (Fig. 3F), excluding the possibility that the transfers described above were caused by the suprathreshold Vernier task alone.

3.4. Double training: Feature training plus location training with an irrelevant but demanding near-threshold task

We know from our previous study that learning can transfer completely to a new location if feature training is accompanied with irrelevant but demanding near-threshold training at the new location (Xiao et al., 2008). Here we examined whether this complete transfer was still specific to the trained orientation. Nine new observers practiced Vernier discrimination at ori1_loc1 and near-threshold contrast discrimination simultaneously using the same Vernier stimulus (see Methods) at ori2_loc2 in alternating blocks for five sessions. Training improved Vernier threshold at ori1_loc1 (MPI = 29.7 ± 2.4%, p < 0.001) and contrast threshold at ori2_loc2 (MPI = 35.9 ± 4.1%, p<0.001). Vernier discrimination at ori1_loc2 also improved (MPI = 31.2 ± 3.0%, p < 0.001), as much as that at trained ori1_loc1 (MPIs at ori1_loc1 vs. ori1_loc2, p = 0.26), showing complete learning transfer. However, learning did not transfer much to orthogonal ori2_loc2 (MPI = 5.0 ± 9.4%, p = 0.30) and ori2_loc1 (MPI = 6.8 ± 4.8%, p = 0.10), showing the same orientation specificity in Fig. 2.

4. Discussion

This study revealed several facts regarding double training that were previously unknown. Active location training is indeed necessary to enable learning transfer across retinal locations. However, if location training is task-irrelevant, as in our previous study (Xiao et al., 2008), learning transfer may be limited to the trained feature orientation only. However, a relevant suprathreshold task at the untrained orientation is sufficient to expand transfer to an untrained orientation, even at the trained location. The latter is especially interesting because it shows that orientation specificity can be decoupled from perceptual learning by a primer at a different retinal location. It is also worth noting that in the current study learning transferred to a diagonal quadrant in the untrained visual hemifield, rather than to the other quadrant in the same hemifield (Xiao et al., 2008). The inter-hemispheric transfer of learning provides additional evidence that perceptual learning occurs in non-retinotopic high-level brain areas.

On the other hand, the orientation specificity of learning transfer with task-irrelevant location training appears to be at odds with our previous TPE training data (Zhang et al., 2010a). The latter shows that foveal learning of orientation discrimination and contrast discrimination can completely transfer to an orthogonal orientation that the observers are passively exposed to in an irrelevant task. This discrepancy is not caused by task differences because the TPE training also enabled nearly full transfer of foveal Vernier learning to an orthogonal orientation (our unpublished data). One possibility is that transfer of peripheral learning in the current study to an orthogonal orientation in the diagonal quadrant involves one extra step of functional connection, in contrast to foveal learning that connects to an orthogonal orientation at the same location. Thus extra processing is required to activate the orthogonal inputs necessary for strengthening the extra functional connections and enabling learning transfer.

In addition, the relationship between the amount of learning transfer and the task demand is less straightforward. Learning appears to transfer more with the task demand in simultaneous double training procedures. But the transfer may already be complete in sequential procedures even with much less demanding suprathreshold location training tasks. The causes of this difference require further investigation. According to our rule-based learning theory, location specificity results from inattention to, or even suppression of visual inputs originating from the untrained location (Zhang et al., 2010a). Location training reactivates these new inputs to connect the learned high-level decision unit to these inputs and enable transfer. It is possible that during simultaneous double training the reactivation by a suprathreshold location training task is not strong enough to fully remove the suppression, but in a sequential procedure the suppression weakens as time passes and the reactivation is now able to remove the remaining suppression.

Recently Sagi (2011) proposed an overfitting theory to explain perceptual learning. This theory posits that when statistically modeling the stimulus representation, an observer through practice could “overfit” the spurious and accidental variations of the stimulus inputs resulting from local noise specific to the trained retinal location and orientation (Mollon & Danilova, 1996), which leads to perceptual learning of the task at hand. But perceptual learning does not transfer to a new location or orientation where the local noise changes. Sagi (2011) suggests that double training enables transfer of learning because “the perceptual modeling process makes use of features shared by all locations and tasks the observer is exposed to during training”, but “overfitting predicts much less learning with two-location as compared with one location training”. This prediction is inconsistent with data shown in Fig. 4 as well as in our previous experiments (Xiao et al., 2008) that double training often produces as much learning as in single training and complete transfer of learning across retinal locations.

Figure 4.

Figure 4

Open in a new tab

Transfer of Vernier learning to a diagonal quadrant trained with an irrelevant near-threshold contrast discrimination task. A. Stimuli for Vernier and contrast double training. For contrast discrimination the observers reported which Gabor had a higher contrast. B. The mean session-by-session Vernier and contrast threshold changes in the trained conditions and pre- and post-training Vernier thresholds in the transfer conditions. C. The MPIs in the trained (left two bars) and untrained transfer conditions (right three bars).

Highlights.

  • Feature-plus-location double training transfers Vernier learning to a diagonal quadrant.

  • Irrelevant location training limits transfer to the trained orientation only.

  • Relevant suprathreshold location training prompts transfer to a new orientation.

  • The amount of transfer depends on the demand of location training and the specific double training procedure.

Acknowledgements

This research is supported by the Natural Science Foundation of China grants 30725018 (CY) and 31000459 (JYZ) and the United States National Institute of Health grants R01-04776 (SAK) and R01-01728 (DML).

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Adini Y, Sagi D, Tsodyks M. Context-enabled learning in the human visual system. Nature. 2002;415:790–793. doi: 10.1038/415790a. [DOI] [PubMed] [Google Scholar]
  2. Ahissar M, Hochstein S. Task difficulty and the specificity of perceptual learning. Nature. 1997;387(6631):401–406. doi: 10.1038/387401a0. [DOI] [PubMed] [Google Scholar]
  3. Bejjanki VR, Beck JM, Lu ZL, Pouget A. Perceptual learning as improved probabilistic inference in early sensory areas. Nat Neurosci. 2011 doi: 10.1038/nn.2796. doi:10.1038/nn.2796. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Crist RE, Kapadia MK, Westheimer G, Gilbert CD. Perceptual learning of spatial localization: specificity for orientation, position, and context. Journal of Neurophysiology. 1997;78(6):2889–2894. doi: 10.1152/jn.1997.78.6.2889. [DOI] [PubMed] [Google Scholar]
  5. Dosher BA, Lu ZL. Perceptual learning reflects external noise filtering and internal noise reduction through channel reweighting. Proc Natl Acad Sci U S A. 1998;95(23):13988–13993. doi: 10.1073/pnas.95.23.13988. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Fahle M. Human pattern recognition: parallel processing and perceptual learning. Perception. 1994;23(4):411–427. doi: 10.1068/p230411. [DOI] [PubMed] [Google Scholar]
  7. Fahle M. Specificity of learning curvature, orientation, and vernier discriminations. Vision Research. 1997;37:1885–1895. doi: 10.1016/s0042-6989(96)00308-2. [DOI] [PubMed] [Google Scholar]
  8. Gal V, Kozak LR, Kobor I, Banko EM, Serences JT, Vidnyanszky Z. Learning to filter out visual distractors. Eur J Neurosci. 2009;29(8):1723–1731. doi: 10.1111/j.1460-9568.2009.06724.x. [DOI] [PubMed] [Google Scholar]
  9. Kahnt T, Grueschow M, Speck O, Haynes JD. Perceptual learning and decision-making in human medial frontal cortex. Neuron. 2011;70(3):549–559. doi: 10.1016/j.neuron.2011.02.054. [DOI] [PubMed] [Google Scholar]
  10. Karni A, Sagi D. Where practice makes perfect in texture discrimination: evidence for primary visual cortex plasticity. Proc Natl Acad Sci U S A. 1991;88(11):4966–4970. doi: 10.1073/pnas.88.11.4966. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Law CT, Gold JI. Reinforcement learning can account for associative and perceptual learning on a visual-decision task. Nat Neurosci. 2009;12(5):655–663. doi: 10.1038/nn.2304. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Mollon JD, Danilova MV. Three remarks on perceptual learning. Spatial Vision. 1996;10(1):51–58. doi: 10.1163/156856896x00051. [DOI] [PubMed] [Google Scholar]
  13. Poggio T, Fahle M, Edelman S. Fast perceptual learning in visual hyperacuity. Science. 1992;256(5059):1018–1021. doi: 10.1126/science.1589770. [DOI] [PubMed] [Google Scholar]
  14. Saarinen J, Levi DM. Perceptual Learning in Vernier Acuity: What is Learned? Vision Research. 1995;35(4):519–527. doi: 10.1016/0042-6989(94)00141-8. [DOI] [PubMed] [Google Scholar]
  15. Sagi D. Perceptual learning in vision research. Vision Research. 2011;51 doi: 10.1016/j.visres.2010.10.019. in press. [DOI] [PubMed] [Google Scholar]
  16. Schoups A, Vogels R, Orban GA. Human perceptual learning in identifying the oblique orientation: retinotopy, orientation specificity and monocularity. J Physiol. 1995;483(Pt 3):797–810. doi: 10.1113/jphysiol.1995.sp020623. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Shiu LP, Pashler H. Improvement in line orientation discrimination is retinally local but dependent on cognitive set. Percept Psychophys. 1992;52(5):582–588. doi: 10.3758/bf03206720. [DOI] [PubMed] [Google Scholar]
  18. Sylvester CM, Shulman GL, Jack AI, Corbetta M. Anticipatory and stimulus-evoked blood oxygenation level-dependent modulations related to spatial attention reflect a common additive signal. J Neurosci. 2009;29(34):10671–10682. doi: 10.1523/JNEUROSCI.1141-09.2009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Teich AF, Qian N. Learning and adaptation in a recurrent model of V1 orientation selectivity. Journal of Neurophysiology. 2003;89(4):2086–2100. doi: 10.1152/jn.00970.2002. [DOI] [PubMed] [Google Scholar]
  20. Tsodyks M, Gilbert C. Neural networks and perceptual learning. Nature. 2004;431(7010):775–781. doi: 10.1038/nature03013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Xiao LQ, Zhang JY, Wang R, Klein SA, Levi DM, Yu C. Complete transfer of perceptual learning across retinal locations enabled by double training. Current Biology. 2008;18(24):1922–1926. doi: 10.1016/j.cub.2008.10.030. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Yu C, Klein SA, Levi DM. Perceptual learning in contrast discrimination and the (minimal) role of context. Journal of Vision. 2004;4(3):169–182. doi: 10.1167/4.3.4. [DOI] [PubMed] [Google Scholar]
  23. Zhang JY, Zhang GL, Xiao LQ, Klein SA, Levi DM, Yu C. Rule-based learning explains visual perceptual learning and its specificity and transfer. Journal of Neuroscience. 2010a;30(37):12323–12328. doi: 10.1523/JNEUROSCI.0704-10.2010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Zhang T, Xiao LQ, Klein SA, Levi DM, Yu C. Decoupling location specificity from perceptual learning of orientation discrimination. Vision Research. 2010b;50(4):368–374. doi: 10.1016/j.visres.2009.08.024. [DOI] [PubMed] [Google Scholar]
  25. Zhaoping L, Herzog MH, Dayan P. Nonlinear ideal observation and recurrent preprocessing in perceptual learning. Network. 2003;14(2):233–247. [PubMed] [Google Scholar]

RESOURCES