NoteERP evidence for distinct mechanisms of fast and slow visual perceptual learning
Introduction
Perceptual learning (PL) represents one kind of skill learning whereby relatively permanent and consistent changes in perception take place with repeated practice or experience. PL involves consolidation of implicit memory formed by training (Squire, 2004), and its time course has drawn a great deal of attention (Alain et al., 2007, Atienza et al., 2002, Karni and Sagi, 1993, Mednick et al., 2005, Tremblay et al., 1998, Yotsumoto et al., 2008). It has been shown that a naïve subject's performance on a simple discrimination task can be significantly improved with only a few trials, a process referred to as fast learning (e.g., Poggio, Fahle, & Edelman, 1992). This fast, within-the-first-session learning is followed by relatively slow learning that accumulates across many training sessions and training days. One important finding about slow learning is that PL may even occur between sessions when no actual training is conducted (Karni & Sagi, 1993). A time period of 1 h immediately after training (Seitz et al., 2005) and various amounts of post-training sleep (Aeschbach et al., 2008, Karni et al., 1994, Matarazzo et al., 2008, Mednick et al., 2003, Stickgold et al., 2000a, Stickgold et al., 2000b) seem to be crucial for the learning to take place and to consolidate. Once acquired, the learning effect can last for a long time, from months to even years without further training (Karni and Sagi, 1993, Polat and Sagi, 1994, Polat et al., 2004, Roth et al., 2005).
From these earlier findings, an interesting question arises: whether both fast and slow learning contribute to the long-term preservation of the PL effect? Although it has been proposed that slow learning may subserve the long-term retention of some perceptual skills (Karni & Sagi, 1993), direct evidence supporting this hypothesis is lacking. Because the effects of fast and slow learning may be intermixed in behavioral measures, it is difficult to address this issue using traditional behavioral measures (i.e., psychophysical threshold, reaction time, accuracy). On the other hand, the event-related potentials (ERPs) that offer high temporal resolution with reasonable spatial resolution may provide additional information to differentiate fast and slow learning contributions to PL.
In the domain of visual research, a number of studies have used ERP and other brain imaging techniques (e.g., functional magnetic resonance imaging, fMRI) to investigate the neural mechanisms of visual PL under different time scales, from tens of minutes (Ding et al., 2003, Itier and Taylor, 2004, Scott et al., 2006, Skrandies et al., 2001, Song et al., 2005, Song et al., 2007, Wang et al., in press) to a few days (Schiltz et al., 1999, Sigman et al., 2005, Yotsumoto et al., 2008). However, to date no published study has been able to examine both the fast and slow PL-associated neural changes in the same experiment, therefore, studies directly comparing the neural mechanisms of fast and slow visual PL are still lacking. Moreover, previous studies did not provide data demonstrating the reliability of specific brain activity after training had been discontinued for months. For these reasons, it is not clear whether neural changes associated with fast or slow PL can be preserved for a long time as the changes in perceptual ability do.
In the present study, by observing the time course of learning-associated changes in different ERP components over a period of six months, we examined whether fast and slow visual PL involves differential ERP changes as well as the roles of fast and slow learning in the long-term preservation of PL. We found that fast and slow learning were reflected by different ERP changes, and only the ERP changes associated with slow learning were retained for a long period of six months. The present study provides, to our knowledge, the first electrophysiological evidence for long-term human adult brain plasticity induced by PL. A general model for perceptual and skill learning was proposed based on these findings and the literature.
Section snippets
Subjects
Ten healthy right-handed adults (21–25 years old, six male) participated in this experiment as compensated volunteers. All had normal or corrected-to-normal vision and were naïve, with no prior experience in the task. Informed consent was obtained from each subject.
Stimuli, task and procedure
A simple visual task reported in our previous studies (Ding et al., 2003, Song et al., 2005) was adopted in the present experiment. There were five stimulus patterns (Fig. 1a), each consisting of four 3.5° × 0.2° lines forming a 2 × 2
Behavioral results
Response accuracy was high (averaged 98%) and stable throughout the training and test sessions. After three consecutive training sessions, RT decreased from 558 ms in S1B1 (i.e., Block 1 in session 1) to 423 ms in S3B3 (Fig. 2a; one-way ANOVA, F(8,72) = 10.20, p = 0.001). Significant decreases in RT were observed not only between the first two blocks in session 1 (S1B1 vs S1B2: decrement 56 ± 25 ms, mean ± S.E.; t(9) = 2.28, p = 0.04), but also across two adjacent training sessions (S1B3 vs S2B1: decrement 15 ±
Discussion
The present study showed that, using RT as an index, visual skill training induced an improvement in behavioral performance not only during training, but also after training. The learning effect was largely preserved even after six months. These results are consistent with previous studies using threshold or accuracy as a measure, supporting the prevailing notion that skill learning is a multistep process continuing beyond the actual training experience (Karni and Bertini, 1997, Stickgold, 2005
Acknowledgments
This work was supported by the National Nature Science Foundation of China grants (30570605, 30600180, 30621004, 90820307), the Humanities and Social Sciences Foundation from the Ministry of Education of China (09YJAXLX025 and 09YJCXLX029), Sun Yat-Sen University 985 Projects (16000-3253187, 2006-90015-3272210), the Ministry of Science and Technology of China grants (2005CB522800), and the Knowledge Innovation Program of the Chinese Academy of Sciences. We thank Prof. Wu Li, Silu Fan, Dr. Ming
References (46)
- et al.
A link between perceptual learning, adaptation and sleep
Vision Research
(2006) - et al.
Benefits of efficient consolidation: Short training enables long-term resistance to perceptual adaptation induced by intensive testing
Vision Research
(2008) - et al.
Progressive N170 habituation to unattended repeated faces
Vision Research
(2006) - et al.
Dissociating top-down attentional control from selective perception and action
Neuropsychologia
(2001) - et al.
Effects of repetition learning on upright, inverted and contrast-reversed face processing using ERPs
Neuroimage
(2004) - et al.
Learning perceptual skills: Behavioral probes into adult cortical plasticity
Current Opinion in Neurobiology
(1997) - et al.
Neuronal mechanisms of perceptual learning: changes in human brain activity with training in orientation discrimination
Neuroimage
(1999) - et al.
Top-down reorganization of activity in the visual pathway after learning a shape identification task
Neuron
(2005) - et al.
Perceptual learning: Psychophysical thresholds and electrical brain topography
International Journal of Psychophysiology
(2001) - et al.
Neural substrates of visual perceptual learning of simple and complex stimuli
Clinical Neurophysiology
(2005)
Memory systems of the brain: A brief history and current perspective
Neurobiology of Learning and Memory
Multi-component correlate for lateral collinear interactions in the human visual cortex
Vision Research
Backward masking suppresses collinear facilitation in the visual cortex
Vision Research
Different dynamics of performance and brain activation in the time course of perceptual learning
Neuron
A role for non-rapid-eye-movement sleep homeostasis in perceptual learning
The Journal of Neuroscience
Changes in auditory cortex parallel rapid perceptual learning
Cerebral Cortex
The time course of neural changes underlying auditory perceptual learning
Learning & Memory
An electrophysiological comparison of visual categorization and recognition memory
Cognitive, Affective, & Behavioral Neuroscience
Early-vision brain responses which predict human visual segmentation and learning
Journal of Vision
Neural mechanisms for visual memory and their role in attention
Proceedings of the National Academy of Sciences of the United States of America
Specificity and generalization of visual perceptual learning in humans: An event-related potential study
Neuroreport
The neurobiology of consolidations, or, how stable is the engram?
Annual Review of Psychology
Introduction
Cited by (34)
Non-invasive neurophysiological measures of learning: A meta-analysis
2019, Neuroscience and Biobehavioral ReviewsAuditory temporal perceptual learning and transfer in Chinese-speaking children with developmental dyslexia
2018, Research in Developmental DisabilitiesCitation Excerpt :Therefore, slow learning is probably a consolidation process of information (Atienza et al., 2002; Galván & Weinberger, 2002; Karni & Sagi, 1993) built upon neural refinement and reorganization of perceptual representations (Recanzone et al., 1993; Zohary, Celebrini, Britten, & Newsome, 1994). Previous studies have revealed that it is the slow learning which promotes the long-term retention of perceptual learning (Karni & Sagi, 1993), involving plastic changes in both lower sensory cortex and higher cognitive regions (Qu et al., 2010). Furthermore, we also observed a loss of correlation between slow learning effect and its transfer to reading fluency, after partialling out the individual performances on executive function.
The effect of badminton training on the ability of same-domain action anticipation for adult novices: Evidence from behavior and ERPs
2017, Neuroscience LettersCitation Excerpt :However, previous cross-sectional studies [10–12] did not observe the training effect of N2. A possible explanation may be related to the duration of the training [53,54]. For example, Song et al. [53] compared the ERP difference of two groups completing short learning (1.5 h) and long learning (3 consecutive days, 1.5 h each day), respectively.
Revisiting the earliest electrophysiological correlate of familiar face recognition
2017, International Journal of Psychophysiology
- 1
These authors contributed equally to this paper.