Relieving the attentional blink in the amblyopic brain with video games

Abstract

Video game play induces a generalized recovery of a range of spatial visual functions in the amblyopic brain. Here we ask whether video game play also alters temporal processing in the amblyopic brain. When visual targets are presented in rapid succession, correct identification of the first target (T1) can interfere with identification of the second (T2). This is known as the "attentional blink". We measured the attentional blink in each eye of adults with amblyopia before and after 40 hours of active video game play, using a rapid serial visual presentation technique. After videogame play, we observed a ~40% reduction in the attentional blink (identifying T2 200 ms after T1) seen through the amblyopic eye and this improvement in performance transferred substantially to the untrained fellow sound eye. Our experiments show that the enhanced performance cannot be simply explained by eye patching alone, or to improved visual acuity, but is specific to videogame experience. Thus, videogame training might have important therapeutic applications for amblyopia and other visual brain disorders.

Figures

Figure 1. The attentional blink test.

We applied a standard attentional blink test to evaluate temporal attention. A sequence of letters was displayed one by one in rapid serial visual presentation. The test consisted of two visual tasks: first, to identify a white letter (T1, randomly selected from 25 uppercase alphabets from A to Z with “X” excluded; letter “H” in this example); second, to detect the presence or absence of a letter “X” (T2, response = yes/no; “yes” in this example) in the letter sequence. The target 2 was presented with a probability of 0.5, at a randomly assigned temporal position after the appearance of T1 (lag = 100–800 ms; Lag 100: “X” was the 1st letter to appear at 100 ms after the onset of T1; Lag 800: “X” was the 8th letter to appear at 800 ms after the onset of T1). A small number of random letters were appended to the end of the sequence. Each letter cycle was 100 ms (1 letter frame, 17 ms + 5 blank frames, 83 ms).

Figure 2. T1 detection.

(a) T1 Detection in the amblyopic eye. T1 detection accuracy was defined as the probability of correctly detecting T1, p(T1); it is also expressed in percentage throughout the main text. The letter targets displayed to observers were suprathreshold in size, 4× – 20× ≥ acuity threshold. The high hit rates reflect that the visual targets were highly visible to the amblyopic eyes in both treatment and control groups. This figure illustrates the visual acuity characteristics of each participant. (b) Comparison of T1 accuracy between the amblyopic eye and the fellow sound eye. Most data points are clustered tightly around the grey unity line, and no significant difference in accuracy was found between the two eyes in each group. Our attentional task was not limited by the reduced visual acuity in the amblyopic eye. (c) The effects of videogame training and occlusion therapy on T1 detection in the amblyopic eye. No significant change in T1 accuracy was found after the intervention in each group.VG: videogame training group. OT: occlusion therapy (patching) group. VAi: isolated letter acuity.

Figure 3. Effects of videogame play on attentional blink.

(a) T2 detection accuracy as a function of time lag. T2 detection accuracy was defined as the conditional probability of the correct detection of T2, given T1 was correctly identified. Open symbols: pre-training performance. Solid symbols: post-training performance. (b) T2 data for each individual observer. Left column: Lag 100–300. Middle column: Lag 400–500. Right column: Lag 600–800. Solid symbols: the lag condition at which the accuracy difference between pre- and post-training measurements reached a 0.05 significance level. White area represents improved accuracy after videogame training. (c) The effect of eye patching on T2 detection performance. In a control experiment, ten participants received occlusion therapy and no significant change in performance was found. AE: amblyopic eye. NAE: non-amblyopic eye.

Figure 4. Improving amblyopic vision with videogame training.

(a) After 40 hours of video-game play (solid symbols), visual acuity improved significantly. In contrast, no significant change in acuity was observed in those participants in the patching group (open symbols). Red symbols: crowded letter acuity. Blue symbols: isolated letter acuity. (b) Two-stage training paradigm. One amblyopic participant (squares in Fig. 4a) practiced a Vernier acuity task, as illustrated in the inset, until acuity improvement reached a plateau. The visual task was to identify the misaligned pair of Gabor patch groupings out of three choices (top, middle or bottom; in the example, the bottom pair is misaligned). Each grouping consisted of 8 Gabor patches. (c) A dramatic boost of T2 detection performance was found at Lag 200 after stage 2 videogame training, but not after stage 1 Vernier acuity training.

Figure 5. The role of visual attention in the normalization of visual acuity.

(a) Acuity improvement as a function of T2 accuracy improvement. The data of observer AS are included here for comparison. After phase one positional acuity training, this observer showed a substantial improvement in visual acuity, but no change in attentional performance was observed (open gray circle). After phase two videogame training, this observer did not show any further acuity improvement, but instead a dramatic boost in T2 accuracy was obtained (solid gray circle). (b) Correlation between spatial attention and temporal attention. A visual counting task was applied to examine how many locations in the visual field the brain can direct attention. A number (1–10) of black circular dots was then presented for 200 ms against a gray background. The dots were randomly positioned in 10 × 10 square cells. Detailed procedures can be found elsewhere.