Perceptual Learning beyond Perception: Mnemonic Representation in Early Visual Cortex and Intraparietal Sulcus

Abstract

The ability to discriminate between stimuli relies on a chain of neural operations associated with perception, memory and decision-making. Accumulating studies show learning-dependent plasticity in perception or decision-making, yet whether perceptual learning modifies mnemonic processing remains unclear. Here, we trained human participants of both sexes in an orientation discrimination task, while using functional magnetic resonance imaging (fMRI) and transcranial magnetic stimulation (TMS) to separately examine training-induced changes in working memory (WM) representation. fMRI decoding revealed orientation-specific neural patterns during the delay period in primary visual cortex (V1) before, but not after, training, whereas neurodisruption of V1 during the delay period led to behavioral deficits in both phases. In contrast, both fMRI decoding and disruptive effect of TMS showed that intraparietal sulcus (IPS) represented WM content after, but not before, training. These results suggest that training does not affect the necessity of sensory area in representing WM information, consistent with the sensory recruitment hypothesis in WM, but likely alters the coding format of the stored stimulus in this region. On the other hand, training can render WM content to be maintained in higher-order parietal areas, complementing sensory area to support more robust maintenance of information.SIGNIFICANCE STATEMENT There has been accumulating progresses regarding experience-dependent plasticity in perception or decision-making, yet how perceptual experience moulds mnemonic processing of visual information remains less explored. Here, we provide novel findings that learning-dependent improvement of discriminability accompanies altered WM representation at different cortical levels. Critically, we suggest a role of training in modulating cortical locus of WM representation, providing a plausible explanation to reconcile the discrepant findings between human and animal studies regarding the recruitment of sensory or higher-order areas in WM.

Keywords: TMS; fMRI decoding; perceptual learning; working memory.

Figure 1.

Procedure and tasks for Experiment 1. A, Participants completed four phases in Experiment 1, including (1) a 2-d pretest, (2) a 6-d training, (3) a 2-d posttest I and (4) a 2-d posttest II. Each test phase comprised of a behavioral and a scanning session, as separately denoted by black and gray bars. The first session of each test phase (behavioral session) included both short-delay and long-delay tasks. The second session of each phase (scanning session) included the long-delay task to address neural activity during WM delay. Training phase used short-delay task only. B, Trial sequence of short-delay (left) and long-delay (right) orientation discrimination tasks. Participants viewed two sequentially presented stimuli and reported whether the test stimulus was tilted clockwise or counterclockwise relative to the sample stimulus in both tasks.

Figure 2.

Procedure and tasks for Experiment 2. A, Participants completed four phases in Experiment 2, including (1) fMRI localizer scans for ROIs definition (V1 and IPS), (2) a 2-d pretest, (3) a 6-d training, (4) a 2-d posttest. Pretest and posttest were completed 1 d before and after the training phase. The two sessions of each test phase were behavioral and TMS sessions, respectively. Both sessions included the long-delay orientation discrimination task. Training phase included the same short-delay task as that used in Experiment 1. B, Trial sequence of the long-delay orientation discrimination task. Participants viewed two sequentially presented stimuli and reported whether the test stimulus was tilted clockwise or counterclockwise relative to the sample stimulus. An online 10-Hz rTMS (five pulses synchronized with 1500 ms after the offset of the sample stimulus) was delivered to one of the stimulation conditions (i.e., V1, IPS, or sham).

Figure 3.

Behavioral results of Experiment 1. A, Participants' discrimination threshold decreased significantly over training sessions. B, Participants' performance during the posttest phases. Left, MPI in the short-delay task during the posttest phase I and phase II. MPI showed learning specificity for the trained compared with the untrained orientation presented at the trained versus untrained location in both posttest phases. Orientation was abbreviated as ori for the condition labels. Right, MPI in the long-delay task during the posttest phase I and phase II. MPI was significantly higher for the trained versus untrained orientation in both posttest phases. Error bars represent SEM across participants.

Figure 4.

Time course of z-transformed BOLD activity in V1 and IPS in posttest I. A, Time course of BOLD activity in contralateral and ipsilateral V1. B, Time course of BOLD activity in left and right IPS. The gray and black triangles indicate time points selected for all subsequent analyses of the delay period and the ITI, respectively. Event labels above the x-axis are shown at the corresponding time points to represent the sample (S), delay (D), and test (T) periods, respectively. Each subplot shows the time course of BOLD response for the trained and untrained orientations. Error bars represent SEM across participants.

Figure 5.

MVPA results of Experiment 1. A, Orientation decoding in contralateral and ipsilateral V1 across three test phases. Classification accuracies were above chance only for the pretest phase in both V1 ROIs. B, Orientation decoding in left and right IPS across three test phases. Classification accuracies were above chance only for the posttest phases in left IPS. The dashed lines denote maximal significance threshold across all conditions obtained from the permutation tests. Note that the significance thresholds were shown only for the purpose of visualization. The reported p values were obtained from the permutation tests for each ROI and each condition, and corrected for multiple comparisons using FDR method. Error bars represent SEM across participants.

Figure 6.

TMS results of Experiment 2. A, Discrimination accuracy at the pretest and posttest phases. From left to right, each pair of bars corresponds to behavioral performance before and after training across different TMS conditions (sham, V1, and IPS). Discrimination accuracy was significantly lower for V1 stimulation versus sham condition in both pretest and posttest phases. Discrimination accuracy was significantly lower for IPS stimulation versus sham condition only in the posttest phase. B, RT at the pretest and posttest phases. RTs were significantly shorter in the posttest than in the pretest phases across TMS conditions. Error bars represent SEM across participants.