Abstract
The impact of action video games on reading performance has been already demonstrated in individuals with and without neurodevelopmental disorders. The combination of action video games and posterior parietal cortex neuromodulation by a transcranial random noise stimulation could enhance brain plasticity, improving attentional control and reading skills also in adults with developmental dyslexia. In a double blind randomized controlled trial, 20 young adult nonaction video game players with developmental dyslexia were trained for 15 h with action video games. Half of the participants were stimulated with bilateral transcranial random noise stimulation on the posterior parietal cortex during the action video game training, whereas the others were in the placebo (i.e. sham) condition. Word text reading, pseudowords decoding, and temporal attention (attentional blink), as well as electroencephalographic activity during the attentional blink, were measured before and after the training. The action video game + transcranial random noise stimulation group showed temporal attention, word text reading, and pseudoword decoding enhancements and P300 amplitude brain potential changes. The enhancement in temporal attention performance was related with the efficiency in pseudoword decoding improvement. Our results demonstrate that the combination of action video game training with parietal neuromodulation increases the efficiency of visual attention deployment, probably reshaping goal-directed and stimulus-driven fronto-parietal attentional networks interplay in young adults with neurodevelopmental conditions.
Introduction
In our daily lives, we are often required to pay attention and process several stimuli that appear simultaneously (Sperling 1960) or temporally close to one another. Attentional control, that is the optimal dynamic interplay between goal-directed and stimulus-driven attention, allows us to control where, how, to what, and when we pay attention to stimuli (Corbetta and Shulman 2002).
A typical situation in which we have to pay attention to several stimuli is reading. Indeed, this cognitive ability requires rapid shifts of spatial and temporal attention to select letters and words through attentional control to reduce the impact of spatial and temporal irrelevant information (Facoetti et al. 2000, 2008; Laasonen et al. 2012; Franceschini et al. 2018; White et al. 2019; Ramamurthy et al. 2021). As a consequence of an early ineffective attentional control, children and adults may show a slow and inaccurate reading, leading to a diagnosis of developmental dyslexia (DD; see Vidyasagar and Pammer 2010; Vidyasagar 2019 for reviews) that is the most frequent neurodevelopmental disorder (DSM-5-TR, 2022). In particular, the attentional control of temporal engagement is slower or disrupted in DD (see Badcock and Kidd 2015 for a meta-analysis). Since DD is a multifactorial and heterogeneous disorder, in which specific deficit in auditory and phonological processing and in memory are also present (Menghini et al. 2010; Mascheretti et al. 2018; O’Brien and Yeatman 2021), training for domain-general cognitive abilities, such as attentional control can be considered (e.g. Facoetti et al. 2003; Breznitz et al. 2013; see Peters et al. 2019 for a review).
A particular type of video game, called the action video game (AVG), shares a set of multisensory features, including high speed, a high degree of perceptual, cognitive, and motor load, temporal and spatial unpredictability, and an emphasis on peripheral processing (Green et al. 2010). Several studies showed that habitual AVG players have better multisensory attentional skills than nonvideo-game players (Green and Bavelier 2003; Green and Bavelier 2007; Green et al. 2012; Antzaka et al. 2017; Mancarella et al. 2022). The breakthrough began when it was demonstrated that an AVG training is able to improve attentional performance even in nonvideo gamers, showing a causal link between the AVG play and augmented attentional control (Green and Bavelier 2007; Li et al. 2009; Green et al. 2012; Oei and Patterson 2013; Bejjanki et al. 2014; see Bediou et al. 2018, 2023 for meta-analysis).
Franceschini et al. (2013) showed that 12 h of AVG improved the reading speed and attention abilities of children with DD, as confirmed by other evidence both in shallow and deep orthographies in children with and without DD (Gori et al. 2016; Franceschini, Bertoni, et al. 2017; Franceschini, Trevisan, et al. 2017; Bertoni et al. 2019, 2021; Franceschini and Bertoni 2019; Peters et al. 2021; Pasqualotto et al. 2022; Bertoni et al. 2024; but, Łuniewska et al. 2018; see Franceschini et al. 2015 for review; see Puccio et al. 2023 for a meta-analysis). The reading far-transfer effect of AVGs could be linked to strong reward signals and a dynamic resetting of the goal-directed and stimulus-driven fronto-parietal networks of attentional control, which are the basis of learning (Bavelier et al. 2012; Dye et al. 2009; Bavelier and Green 2019; Zhang et al. 2021; Jordan and Dhamala 2023).
The AVG enhancement of attentional control is seen also in temporal attention (Green et al. 2010). AVG players show a reduced attentional blink (AB) as compared to nonvideo game players (Dye and Bavelier 2010) and after an AVG training in nonvideo game players (Green and Bavelier 2003; Oei and Patterson 2015). AB is a phenomenon in which a transitory impairment of attention occurs when multiple targets have to be processed in close temporal proximity (Raymond et al. 1992; Kranczioch et al. 2003; Snir and Yeshurun 2017). In AB tasks, a rapid serial visual presentation (RSVP) of stimuli is displayed at a single location in which 2 of them are defined as targets presented at various intertarget time intervals (Kranczioch et al. 2003).
During the AB, an amplitude reduction of P300 event-related potential (ERP) suggests a processing interference (Kranczioch et al. 2003; Ronconi et al. 2016a). In fact, P300 amplitude is a measure of the availability of cognitive processing resources for target perception and recognition (Luck 1998; Mishra et al. 2011). AVG players show larger P300 amplitudes than nonvideo game players during an RSVP task (Mishra et al. 2011). Wu et al. (2012) found larger P300 amplitudes after 10 h of AVG training, suggesting a general enhancement of the attentional control plasticity.
A further tool to facilitate brain plasticity is the transcranial electrical stimulation (tES; Contò et al. 2021). Different tESs, such as transcranial direct current stimulation (tDCS), and transcranial magnetic stimulation methods have been used to induce effects on reading skills and as a potential treatment for adolescent and adults with reading difficulties and/or DD (e.g. Turkeltaub et al. 2012; Costanzo et al. 2012, 2013, 2019; Lazzaro et al. 2021; see Turker and Hartwigsen 2021 for a review). The transcranial random noise stimulation (tRNS), in which the stimulation is alternating at a random frequency and amplitude within a specific range (Reed and Cohen Kadosh 2018), has been shown to be the most efficacious neuromodulatory technique to enhance perceptual processing (Fertonani et al. 2011; Romanska et al. 2015; Herpich et al. 2019) when it was applied during task execution (Pirulli et al. 2013), and it has been shown to be the most effective for blinding (Sheffield et al. 2022). Indeed, tDCS and tRNS have similar aftereffects, but tRNS might be better suited to control the placebo effect (Ambrus et al. 2010). tRNS improves the signal-to-noise ratio through repeated amplification of subthreshold oscillatory activity that potentiates task-related neural activity (Fertonani et al. 2011; Miniussi et al. 2013; Pavan et al. 2019), enhancing the response to goal-directed targets (Moret et al. 2021). Delivering tES over an area can cause a cascade of functional changes spreading within a specific network (Krause et al. 2017; Costanzo et al. 2019), or modulate large-scale networks (Kunze et al. 2016; Lema et al. 2021). A single session of parietal tRNS during reading tasks improves reading performance in a large sample of young adults (Bertoni et al. 2023).
In this study, we test the effect of the combination of an intensive but brief AVG training with bilateral tRNS on the posterior parietal cortex in young adults with DD.
We stimulated the posterior parietal cortex which controls visual attention (Corbetta and Shulman 2002) and encodes the spatio-temporal information by integrating sensory signals from multiple modalities (Andersen 1997). In particular, we chose high-frequency tRNS (hf-tRNS), which is capable of inducing excitatory effects and long-term potentiation (Rioult-Pedotti et al. 2000; Terney et al. 2008; Campana et al. 2016), hypothesizing a boost of the AVG training effect after this new methodological approach. The effect of active hf-tRNS is compared to a sham stimulation for excluding a possible placebo effect elicited by the active stimulation (Brancucci et al. 2023). In addition, we investigated possible variations induced by the training on the P300 through ERP recorded during a temporal attention task. The long-lasting effect in the hf-tRNS group on reading skills was investigated 4 mo after the end of the training. Finally, the relationship between temporal attention and reading enhancements and the individual behavioral enhancements in the hf-tRNS group were tested.
Materials and methods
Participants
The sample size was determined by a priori power calculation conducted using GPower (Erdfelder et al. 1996) to estimate the smallest sample size needed to detect an effect size F = 0.36 (see Bediou et al. 2023; Bello et al. 2023 for meta-analyses). The analysis was modeled for a multivariate analysis of the variance (MANOVA), 2 groups with 4 measurements, and a 2-tail alpha equal to 0.05. Under these assumptions, the minimal sample size predicted to be needed with 80% statistical power was 18 participants.
Participants were selected from a sample of about 700 university students. Selection criteria were: (i) their AVG experience: It was acquired through a questionnaire in which we asked them if they played video games; if yes, which video game, how many times a week, and for how long (no use of AVGs for >5 h a week in the last 6 mo was the inclusion criteria); and (ii) the score in the “adult reading history questionnaire” was >0.35 (Lefly and Pennington 2000). During a first evaluation, they were tested on their effective reading skills through consolidated clinical tests for reading skills (Sartori et al. 2007; Judica and De Luca 2005). The study was proposed to participants who obtained a performance ≤−1.5 SD below the average in at least one of the reading tests in accuracy and/or speed (DSM-5-TR, 2022). Twenty-six participants were originally selected based on our criteria. Five participants did not agree to participate, and 1 participant was excluded because of her/him epilepsy syndrome. The final sample of the participants who agreed to participate in the entire rehabilitation study was composed of 20 native Italian speaking adults (11 females and 9 males; mean age = 22.85 yr, SD = 6.53) with DD due to their significant reading difficulties despite their high level of education. This sample was randomly divided into 2 groups by using a block randomization procedure, where 20 individuals were randomized into 2 conditions in blocks of size randomly selected between 2 and 8 (Efird 2011). A group received active hf-tRNS stimulation (hf-tRNS group; mean age = 22.1, SD = 4.65; 7 females) and the other one received placebo stimulation (sham group; mean age = 23.6, SD = 8.19; 4 females) during the AVG training. The analysis did not show significant differences in age (P > 0.621) and female/male distribution (P > 0.178). All the participants gave written informed consent after a description of the study in accordance with the principles of the Declaration of Helsinki. The study was approved by the Ethics Committee of the Department of General Psychology, University of Padova (Protocol number: 1551).
Procedure
Participants were individually tested and trained in a dimly lit and quiet room. Reading and temporal attention skills of the participants were tested between 3 and 5 days before the start of the training (pretraining) and were retested between 3 and 5 days after the end of training programs (post-training), with an interval of at least 24 h after the final play session to ensure that any transient effects of game-play were removed as potential confound (Green and Bavelier 2015). In the pre- and post-training evaluations, electrophysiological data were also recorded during the AB task execution.
Training consisted of 12 days of AVG sessions of 75 min each for a total of 15 h. During the AVG play, the participants sat at a distance about 150 cm from a 27-in TV screen. A commercial PlayStation 3 video game from Activision (deemed suitable for people aged ≥18 by the Pan European Game Information) called as “Call of Duty: Modern Warfare 2” was used, which is 1 of the most prevalent and successful first-person shooter games (Spence and Feng 2010; Nguyen and Bavelier 2023).
The hf-tRNS stimulation was generated by a Brain Stim stimulator by E.M.S. (Bologna, Italy) and was delivered via a pair of identical rubber scalp electrodes (4 × 4 cm), covered with saline-soaked synthetic sponges. The hf-tRNS (frequency = 101–640 Hz and intensity = 1.5 mA) was delivered over bilateral posterior parietal areas (P3 and P4 according to the 10–20 electroencephalographic [EEG] system: Herwig et al. 2003; Okamoto et al. 2004). Elastic bandages were used to maintain the electrodes in place during the training. Duration of the stimulation was about 20 min. A ramp-up of 30 s was used before the beginning of the training. In the sham stimulation, a ramp-up of 30 s was immediately followed by a ramp-down of 30 s. Participants and experimenters were blinded for stimulation conditions. Indeed, the experimenters that evaluate the cognitive performances pre- and post-training did not treat the participants and vice versa.
Finally, the reading skills of the hf-tRNS group were retested after 4 mo from the end of the training (follow-up; Fig. 1).
Schematic representation of the training project timeline.
Behavioral evaluation
Pseudoword reading task
Phonological decoding abilities were measured using 3 Italian standardized lists of 48 pseudowords each for a total amount of 127 syllables each (Sartori et al. 2007). Reading time (seconds) and number of errors were measured. Participants were invited to read each pseudoword aloud as quickly and as accurately as possible. The reading time was measured from the moment the experimenter turned the sheet until the participant pronounced the last letter of the last pseudoword of the list. A wrongly read pseudoword was counted as 1 error independently from the quantity of wrong letters or syllables pronounced. Self-correction was not counted as an error.
Word reading task
Word reading abilities were evaluated using 2 standardized texts (1 in pre- and 1 in post-training evaluation) of similar length and reading difficulty (Judica and De Luca 2005). Reading time (seconds) and number of errors were measured. Participants were invited to read each text aloud as quickly and as accurately as possible. The reading time was measured from the moment the experimenter turned the sheet until the participant pronounced the last letter of the last word of the text. A wrongly read word was counted as 1 error independently from the quantity of wrong letters or syllables pronounced. Self-correction was counted as a 0.5 error.
Temporal attention task
Temporal attention was evaluated through an RSVP task to measure the AB effect (see Ronconi et al. 2016b for details). The experiment was presented on a cathode ray tube monitor, refreshing at 100 Hz. Stimuli presentation and data acquisition were performed using Presentation (Neurobehavioral Systems, version 14). Testing occurred individually in a dimly lit room and participants sat at a distance of 60 cm. On each trial, after a central fixation point (a cross-shape subtending 2° × 2°) showed for 2,000 m, an RSVP stream was presented containing 22 items (2 targets and 20 distractors) that were presented one after the other in the center of the screen. At the end of each stream, participants were given an unlimited time to report the target letters in the correct order and were required to guess if they were unsure. The target letters could be the 21 letters of the alphabet (chance level about 5%), and both target letters and digit distractors were randomly assigned on each trial with the restriction that items in successive trials could not be the same.
Black distractors were presented for 100 ms. Target letters were presented for 60 ms and were followed by a 40 ms hash (#) mask. In this manner, the target-mask duration was equivalent to the distractor duration. In order to prevent predictability, occurrence of the first red target letter (L1) randomly appeared as the fourth, fifth, or sixth item in the stream. The second black target letter (L2) appeared at lags 3 or 8 with equal frequency (see Fig. 2).
Schematic representation of the AB task. The sizes of the stimuli are not real.
The total number of trials was 230 (divided in 5 miniblocks of 46 trials), comprising 205 L2-present trials and 20 L2-absent trials (catch trials) and 5 initial practice trials.
EEG data acquisition and preprocessing
EEG was recorded both pre- and post-training with 32 active Ag/AgCl electrodes (Micromed System Plus analysis system, Micromed) distributed over the scalp according to the International 10–20 EEG System during the execution of the AB task in hf-tRNS (10 participants) and sham (8 participants) group. The ERP data of 2 participants from the sham group were missing due to corrupted files. Data were sampled online at 512 Hz. All electrodes were online referenced to the right mastoid. Vertical and horizontal electro-oculography were recorded from electrodes placed below each eye and on the 2 external canthi. Offline, data were downsampled to 256 Hz, recomputed to an average reference, band-pass-filtered between 0.5 and 30 Hz using Butterworth filters, and notch-filtered to remove 50-Hz line noise. To analyze the event-related EEG activity elicited by the presentation of L1 (target 1 analysis) and by the presentation of L2 (target 2 analysis), the data were segmented, separately, in epochs ranged between −200 ms and +800 ms relative to the L1 and L2 onset. In both analyses, data were baseline-corrected by subtracting the mean over a 200 ms prestimulus window, and all epochs with incorrect behavioral responses (e.g. L1 correct and L2 incorrect) were rejected. This procedure was adopted, separately, both for lag 3 and lag 8 conditions. We computed the independent component analysis using EEGLAB’s runica function to identify and remove artifacts related to eye-blinks and muscle movements for each subject. Subsequently, absolute amplitude was calculated in each trial, and we rejected all trials that exceeded ±80 mV. Spherical interpolation was carried out on individual bad electrodes if required (average number of interpolated electrodes: M = 2.7; SD = 1.42).
EEG data analysis
Based on previous literature investigating the neuroelectric correlates of AB (e.g. Dell'acqua et al. 2003; Kranczioch et al. 2003, 2005; Mishra et al. 2011), through visual inspection of the ERPs waveforms and topographical scalp maps, we focused our EEG data analysis on a neural hallmark of AB phenomena, the P300 ERP component elicited by L1 and L2 presentations, only in lag 8 condition. We have excluded from the analysis the lag 3 condition because the time between L1 and L2 is too short to analyze the P300 component. In detail, the P300 component was calculated choosing, respectively, time windows of 200–350 and 350–550 ms for L1 and L2 presentations. Electrodes for P300 analyses were selected following previous findings (e.g. Dell'acqua et al. 2003) and following scalp maps distribution, thus including and averaging the neural activity of the following centro-parietal electrodes: P3, Pz, P4, Cp3, Cpz, and Cp4.
Statistical analysis
Between-group comparisons of pretraining performance will be performed through t-test analysis. Then, to understand the effect of the training programs, while analyzing at the same time all considered variables, MANOVA and t-test will be run. If differences are present between the 2 groups in the pretraining analysis, these variables will be used as covariate in the MANOVA, i.e. multivariate analysis of covariance (MANCOVA) to be sure that this difference did not affect the effect induced by the training programs. In addition, correlational analysis will be performed to investigate a possible link between reading and temporal attention skills. Analysis of variance (ANOVA) will be performed to investigate the effect on EEG data. Analyses will be conducted using SPSS software.
Results
Behavioral results
In the pretraining, the analysis did not show significant effects between the hf-tRNS and sham groups in pseudoword reading skills, both in speed and errors (Ps > 0.111) and in the AB performance at both lag conditions (Ps > 0.32). The analysis showed a significant effect between the 2 groups in word reading speed (t(18) = 3, P = 0.008; hf-tRNS mean = −1.98, SD = 0.97; sham mean = −5.66, SD = 3.75), but not in word reading errors (P > 0.472).
The effect of the training on pseudoword and word reading skills was tested considering the reading efficacy, and it was calculated as the ratio between the accuracy (number of pseudowords or words read correctly) and the time (seconds) in order to have a variable that considers accuracy and reading speed contemporary and to avoid the trade-off speed accuracy effect (Green and Bavelier 2015).
Between-group differences in the reading performance and temporal attention enhancements
A MANCOVA with the word text reading speed as covariate with a 4 × 2 design was run. The within-subject factors were the enhancement (pretraining minus post-training scores) obtained in pseudoword lists and word texts reading efficacy, and the enhancement obtained in the AB task at the 2 lag conditions, while the between-subject factor was the group (hf-tRNS vs. sham). The group main effect (F(14) = 3.73, P = 0.029, η2p = 0.52) was significant. For this reason, we have run a within-subject analysis to test if the enhancement obtained between pre- and post-trainings were different from an absence of improvement (no gain = 0) through 1-sample t-tests for the variables analyzed in the MANCOVA.
Pseudoword reading
The t-tests reveal that the enhancement in phonological decoding efficiency of hf-tRNS group was significantly different from 0 (t(9) = 8.65, P < 0.0001). On the contrary, the same analysis on the sham group did not show a significant effect in phonological decoding (t(9) = 1.63, P > 0.14; Fig. 3A; see also Supplementary Results and Fig. S1).
A) Reading efficiency enhancement calculated as difference between pre- and post-training performances. The asterisks indicate the statistical difference other than 0; B) attentional skills enhancement at both lag conditions. The asterisks indicate the statistical difference other than 0. Error bars indicate standard error of the mean.
Word text reading
The t-tests reveal that the enhancement in word reading efficacy of hf-tRNS group was significantly different from 0 (t(9) = 2.52, P = 0.033). On the contrary, the same analysis on the sham group did not show a significant effect in word reading (t(9) = 1.36, P > 0.21; Fig. 3A; see also Supplementary Results and Fig. S2).
Attentional blink
The t-tests reveal that the enhancement in attentional skills of hf-tRNS group was significantly different from 0 at both lag conditions (lag 3: t(9) = 4.88, P = 0.001; lag 8: t(9) = 4.35, P = 0.002). On the contrary, the same analysis on the sham group did not show a significant effect in attentional skills neither to lag 3 (t(9) = 0.61, P = 0.56), nor to lag 8 (t(9) = 1.69; P = 0.13; Fig. 3B; see also Supplementary Results and Fig. S3).
Long-lasting effect of hf-tRNS on reading performance
Four months after the training, the hf-tRNS group maintained the enhancement in pseudoword reading efficiency (post-training vs. follow-up: t(8) = 1.01, P > 0.34), whereas the word reading efficiency enhancement improved even more (post-training vs. follow-up: t(8) = 4.85, P = 0.001).
Relationship between training-induced temporal attention and reading enhancements
Correlation analyses were run on the enhancement obtained in pseudoword and word reading efficiency and in attentional skills (considering the mean between the 2 lag conditions), controlling for word reading speed measured at pretraining. The results showed a positive relation between the enhancements in pseudoword reading efficiency and in attentional skills (r = 0.60, P = 0.007; Fig. 4).
The relationship between the training-induced enhancements in temporal attention and pseudoword reading efficiency.
Individual results of behavioral enhancements in the hf-tRNS group
The aim of this analysis is to better estimate the individual advantage of hf-tRNS in the temporal attention and reading performance enhancements for each participant. The gain obtained in word, pseudoword reading and AB task in each hf-tRNS participant was compared to the mean of the gain obtained in the sham group. The results indicate that the gain in pseudoword reading of 10 hf-tRNS participants (100%) was higher than the mean of the gain of the control group. The gain in word text reading efficiency of 7 hf-tRNS participants (70%) was higher than the mean of the gain of the control group. The gain in AB performance of 9 hf-tRNS participants (90%) was higher than the mean of the gain of the control group.
EEG results
Mean amplitude of the P300 at lag 8
A 2 × 2 × 2 ANOVA on the mean amplitude of P300 at lag 8 was carried out. The time (pre- and post-trainings) and target (L1 and L2) were the within-subjects variables, while the group (hf-t-RNS and sham) was the between-subject variable. The time (F(1,16) = 0.058, P = 0.812, η2p = 0.004), target (F(1,16) = 2.445, P = 0.137, η2p = 0.133), and group (F(1,16) = 1.027, P = 0.326) main effects were not significant. Importantly, time × target × group interaction was significant (F(1,16) = 6.372, P = 0.023, η2p = 0.285; see Fig. 5. Post hoc t-tests were separately executed for L1 and L2 targets. Only in hf-tRNS group, the P300 mean amplitude significantly decreased for L1 target (pre-: M = 1.097, SD = 1.245 and post-training: M = 0.614, SD = 1.299; t(9) = 2.220, P = 0.026), whereas it significantly increased for L2 target (pre-: M = 0.874, SD = 1.084 and post-training: M = 2.119, SD = 1.495; t(9) = −3.453, P = 0.004).
AB ERPs waveforms on centro-parietal electrodes (P3, Pz, P4, Cp3, Cpz, and Cp4) and scalp maps results. The upper panel shows ERPs waveforms and topographic scalp maps elicited by the AB target 1 (L1) presentation during pre- and post-training, separately for hf-tRNS (left panel) and sham (right panel) groups. Topographical activation maps represent the average activity of P300 (200–350 ms) amplitude during pre-, post-training, and post–pre differences. Gray shaded bars indicate the P300 time window on which the statistical analysis was computed. The bottom panel shows ERPs waveforms and topographic scalp maps elicited by the AB target 2 (L2) presentation during pre- and post-training separately for hf-tRNS (left panel) and sham (right panel) groups. Topographical activation maps represent the average activity of P300 (350–550 ms) amplitude during pre-, post-training and post–pre differences. Gray shaded bars indicate the P300 time window on which the statistical analysis was computed.
Peak latency of the P300 at lag 8
The first ANOVA run on P300 peak latency with the aim of testing whether neural activity of L1 was influenced by the time and group factors did not show significant main effects of time (F(1,16) = 0.04, P = 0.996, η2p = 0.001), group (F(1,16) = 0.450, P < 0.512, η2p = 0.027), nor their interaction (F(1,16) = 0.062, P < 0.806, η2p = 0.004). The ANOVA performed on P300 peak latency with the aim of testing whether neural activity of L2 was influenced by the time and group factors did not show significant main effects of time (F(1,16) = 1.410, P = 0.252, η2p = 0.081), group (F(1,16) = 3.211, P < 0.092, η2p = 0.167), nor their interaction (F(1,16) = 0.534, P < 0.475, η2p = 0.032).
Discussion
The present study was designed to compare the effect of an intensive but brief AVG training combined with hf-tRNS or with sham stimulation in adults with DD. In particular, we tested the behavioral and electrophysiological effects of a bilateral hf-tRNS on the posterior parietal cortex, administered during an AVG training, on reading and temporal attention performance.
The results show that only coupling AVG training with hf-tRNS enhances words and pseudowords reading efficiency and visual temporal attention mechanisms. In particular, only AVG with the bilateral hf-tRNS on the parietal cortex enhances the optimal deployment of temporal attention on 2 visual targets. Accordingly, only after the AVG combined with parietal hf-tRNS, the ERP data show an increased P300 amplitude of the temporal attention deployment at the second visual target, associated with a decreased P300 amplitude of the temporal attention capture at the first salient visual target. These specific changes of the P300 amplitude suggest that both reading and temporal attention enhancements could be obtained by reshaping goal-directed and stimulus-driven fronto-parietal attentional networks interplay (see Corbetta and Shulman 2002 for a review). In particular, our P300 amplitude results could indicate a decreased activation of stimulus-driven attentional control, combined with an increased activation of goal-directed attentional control. These findings suggest that attentional control enhancement induced by the combination of bilateral hf-tRNS on the posterior parietal lobes and AVG training in the developed brain involves an optimal distribution of processing resources rather than a general neuroplastic enhancement of processing resources (Wu et al. 2012; see Von Bastian et al. 2022 for a recent review). In particular, an increase in functional connectivity within the goal-directed and stimulus-driven attention networks after posterior parietal cortex stimulation could be the neurobiological basis of this optimal distribution of processing resources. Indeed, the changes in functional connectivity within attentional neural networks are a marker for the enduring effect of tRNS upon behavior (e.g. Contò et al. 2021). In addition, these findings could partially confirm the functional interactions within and between the stimulus-driven and the goal-directed attentional control networks in video game players (Jordan and Dhamala 2023) and also suggest a causal link in the interplay of these 2 attentional networks for reading and attention enhancements also in atypical developed brain.
The increase in the P300 amplitude recorded during the execution of the temporal attention task, which is in line with the results of Mishra et al. (2011). Indeed, since the P300 wave amplitude is linked to the ability to inhibit processing of distractors (Wu et al. 2012), this means that AVG combined with bilateral hf-tRNS on the posterior parietal lobes enhances the attentional control and the ability to resist distraction.
Although further studies with larger samples are needed, individual data analysis appears to indicate that the reading and attentional enhancements induced by this new methodological approach of hf-tRNS and AVG training combination, impacted on a large part of adults with DD of our small sample despite the neural plasticity in developed brains is dramatically reduced (Bavelier et al. 2010). To note that the individual data analysis in larger samples of children with DD treated with AVG has shown similar reading and attentional enhancements (see Puccio et al. 2023 for a meta-analysis).
Moreover, these far-transfer reading effects induced by an efficient attentional control training are long lasting because after 4 mo from the end of the combined training, the enhancement of the phonological decoding efficiency was stable and the word reading efficiency kept improving.
These results show the ability of hf-tRNS to boost the effect of AVG training, as demonstrated in other studies with perceptual learning paradigms (Cappelletti et al. 2013). In addition, our data demonstrate that coupling an intensive but brief AVG training with hf-tRNS to critical brain regions for attentional control, such as the posterior parietal cortex, resulted in long-lasting improvement transferable on an untrained ability, such as reading, that shares common neurocognitive components with visual attention.
Indeed, the correlational analysis, running on the entire sample, shows that there is a positive relation between the enhancement in the ability to optimally deploy temporal attention on both target letters in the AB task and the enhancement in phonological decoding efficiency. These results suggest that the improvements in reading efficiency and in temporal attention are causally linked. In detail, the relation between the enhancement of the accuracy to detect both target letters and the reading skill could underline the importance to improve the temporal attention processing to better filter the relevant from irrelevant information for enhancing reading skill. These findings confirm the evidence about the link between reading skills and visuo-spatial attention found in children with and without DD (Puccio et al. 2023 for a meta-analysis; Pasqualotto et al. 2022). However, our results suggest that the link between reading and attention involves not only spatial but also the temporal domain.
The temporal attention enhancement at both lag conditions suggest that participants become more efficient to suppress the distraction and faster and more faithful to extract task-relevant information, as pointed out by Green and Bavelier (2015) and recently demonstrated by Zhang et al. (2021).
Our results confirm that an AVG training induces transfer effect (Green and Bavelier 2015) and that this effect is most strongly present when we associate cognitive training (i.e. AVG) with brain stimulation (i.e. tRNS; Cappelletti et al. 2013). These evidence is in line with Cappelletti et al. (2013) in which with a bilateral tRNS on posterior parietal lobe (i.e. P3 and P4), administered for 5 consecutive days, 20 min each, during a numerosity discrimination task, induced an enhancement in the trained numerosity task that is retained up to 4 mo after the training. Furthermore, the improvement was transferred to untrained abilities that share both cognitive and anatomical resources with the numerosity task, such as time and space processing (Cappelletti et al. 2013).
The issue of transfer enhancements on untrained skills is fundamental in clinical and educational contexts where the principal aim is to induce effects in the most generalizable possible way (Cappelletti et al. 2013; see Green and Bavelier 2012 for a review). The ability to generalize the effect to untrained context must be the rationale of cognitive training programs (Von Bastian et al. 2022). As mentioned by Von Bastian et al. (2022), the transfer to untrained tasks or every day functioning is present in training interventions targeting processing speed in which the participants have to react as quickly and as accurately as possible to a stimulus, often in combination with attentional control demands, as happens during AVG playing. In addition, since 1 of the aspects that determine the success of the transfer is how close the trained and untrained skills are (Cappelletti et al. 2013), this means that the functioning of the fronto-parietal networks, involved during the AVG training, is fundamental for both attention and reading performance (Bertoni et al. 2023).
Our results should be confirmed by future training studies that better matched the participants between different training groups. Another possible limit of our study is the absence of a group untrained with AVG, but the presence of 2 groups in which the AVG training is the same, brings a methodological advantage linked to the blindness of the participants. In fact, it is difficult to maintain the blindness of the participants, with respect to the experimental condition, in studies in which the effectiveness of an intervention is evaluated (Simons et al. 2016). Consequently, there is always the possibility that it is not the content of the intervention itself to bring improvements, but the expectation of the participants (Green and Bavelier 2015). In this case, the participants could not have different expectations related to the training. However, it should not be neglected that in the AVG training, the evidence suggests that participant expectation cannot explain the effects observed (Green and Bavelier 2015; Zhang et al. 2021). We conclude that the bilateral hf-tRNS on the posterior parietal cortex used as a boost for an intensive but brief AVG training has strong potential to augment temporal attention and reading skills in adults with DD.
Some studies have demonstrated that only 10–20 h of AVG training was able to significantly change the AB performance in healthy adults (e.g. Green and Bavelier 2003; Oei and Patterson 2013). By contrast, other studies have demonstrated that 40 h are necessary to change visual and reading skills in clinical samples, such as amblyopic adults (e.g. Li et al. 2015; Vedamurthy et al. 2015). Since our sample was composed of 20 university students with DD, 15 h with only AVG training was probably too short to elicit enhancements on these neurocognitive skills. Further studies are necessary to investigate the optimal duration of AVG training to improve temporal attention and reading skills in young adults with DD.
In sum, our results show long-lasting plastic changes when an AVG training is coupled with a neuromodulatory intervention. Although scientific evidence seems to demonstrate the specificity of the effects of AVGs on attention and reading (see Bediou et al. 2023 and Puccio et al. 2023 for meta-analyses), other types of video games, for example, sports and race games, considered action-like, could also be effective. Further studies are therefore necessary to demonstrate whether other types of video games combined with tRNS can induce similar effects on attention and reading and to find the most appropriate and efficacious parameters for applying this paradigm to clinical practice.
Acknowledgments
We thank all participants who took part in the study.
Author contributions
Sara Bertoni (Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Software, Writing—original draft), Sandro Franceschini (Conceptualization, Investigation, Methodology, Software, Writing—review & editing), Martina Mancarella (Investigation, Writing—review & editing), Giovanna Puccio (Investigation, Writing—review & editing), Luca Ronconi (Conceptualization, Data curation, Methodology, Software, Writing—review & editing), Gianluca Marsicano (Data curation, Formal analysis, Writing—review & editing), Simone Gori (Conceptualization, Methodology, Writing—review & editing), Gianluca Campana (Conceptualization, Funding acquisition, Methodology, Project administration, Resources, Writing—review & editing), and Andrea Facoetti (Conceptualization, Formal analysis, Funding acquisition, Methodology, Project administration, Resources, Writing—review & editing).
Funding
This work was funded by a grant from MIUR (Dipartimenti di Eccellenza DM 11/05/2017 n.262) to the Department of General Psychology, University of Padua, to SF, the CARIPARO Foundation (Borse di Dottorato CARIPARO 2015 to SB), Progetto MIUR Dipartimenti di Eccellenza (DM 11/05/2017 n. 262) “Use-Inspired Basic Research—Un modello innovativo per la ricerca e la formazione In Psicologia” CUP: C96C18000450005 to SB and GP, and Ateneo Research Project STaRs (Supporting Talented Researchers) azione 1 assegni di ricerca anno 2020 CUP F54I19000980001 a budget dell”Ateneo to SB, Research Foundation-Flanders (FWO) under Grant 1112023N to MM, and MUR PRIN Project CUP:F53D23004610006—ID MUR: 2022772HTJ_01 to S.G. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Conflict of interest statement: None declared.
