New Insights Into Slow Reading: Behavioral and Electrophysiological Analyses of Visual Word Recognition in Arabic

Abstract

This study investigated the behavioral and electrophysiological differences between fast and slow readers among Arabic-speaking university students. We employed a classification methodology similar to that used in the rate versus accuracy approach of dyslexia subtyping. Fifty-five native Arabic-speaking university students participated in a lexical decision task involving high-frequency (HF) words, low-frequency (LF) words, and pseudowords (PWs). Behavioral and event-related potentials (ERPs) data were collected during the task. Participants were categorized as fast or slow readers based on their mean reaction time (RT) across all conditions, with those with RTs below the 65th percentile classified as fast readers and those with RTs above the 75th percentile classified as slow readers. Behaviorally, we observed a frequency effect, with faster RTs for HF words compared to LF words and PWs. At the electrophysiological level, we found a reader effect on the latency of the early ERP components (N170, P2, N2, and P3), with earlier peak latencies for fast readers. Additionally, the P600 component showed a larger amplitude and earlier peak for HF words compared to LF words and PWs. Fast readers exhibited a larger P600 amplitude and an earlier P600 peak for HF words compared to slow readers. These findings provide novel insights into word recognition processes in Arabic-speaking adults, shedding light on the neural mechanisms underlying differences between fast and slow readers. The results are discussed in relation to previous research on word recognition in both typical and dyslexic readers.

Keywords: P600 component; accuracy‐based dyslexia; component amplitude; event‐related potentials (ERPs); visual word processing; word frequency.

FIGURE 1

Graphs illustrating the statistical p values (green for p < 0.01; dark red for the highest p < 0.001 values) of the pointwise t‐test analysis (see Section 2) performed on all time points from 0 to 800 ms (x axis) and all 64 recording sites (y axis) for the comparisons (A) HF words vs LF words and (B) HF words vs PWs. These graphs show that robust and consistent differences appeared in the P600 time window. The insets in the upper left (A₁ and B₁) illustrate t maps for the time point at 550 ms (also see color scales for t values). The inset maps (lower left A₂ and B₂) illustrate the location of the electrodes depicting significant differences during these time points, and show that in HF words versus LF words and HF words versus PWs, a high number of significant and adjacent electrodes were observed. Note that these insets are located for aesthetic purposes within the graphs where no significant differences were found.

FIGURE 2

(A) Superposition of the grand‐mean ERP (with standard error of the mean [SEM]) traces for the three word conditions (black: HF word; red: LF word; and green: PW) from the left (electrodes P5, P7, and PO7) and right (electrodes P6, P8, and PO8) posterior regions of interest. (B) Superposition of the grand‐mean ERP traces from the same regions of interest for the average ERPs (with SEM) across conditions as a function of reader type (black: fast readers; red: slow readers). The traces, enlarged from 0 to 500 ms poststimulus (see inset for electrodes' location), exhibit the successive P1, N170, P2, N2, and P3 components (see arrows) in the different conditions. (C) Graph illustrating box‐and‐whisker plots representing the distribution of ERP amplitudes for each early component (P100, N170, P2, N2, and P3). For each component, two box plots are displayed: one for the left hemisphere (black) and one for the right hemisphere (gray). The box plots (here and in the other graph) illustrate the median, mean, interquartile range (IQR), and range of the amplitude data for each component and hemisphere. Outliers (here and in the other graph) are not explicitly marked but are contained within the range. (D) Box‐and‐whisker plots representing the distribution of ERP latencies for each early component (P100, N170, P2, N2, and P3). For each component, two box plots are displayed: one for fast readers (black) and one for slow readers (red).

FIGURE 3

(A) Superposition of the grand‐mean ERP traces (with SEM, from 0 to 800 ms) for the three language conditions (black: HF words; red: LF words; and green: PWs) from three left, three midline, and three right centro‐parietal sites (as demonstrated also by the inset in B) selected on the basis of pointwise t‐tests (Figure 1) and which best exhibited the ERP P600 component's differences. (B) The superposition of mean P600 signal computed from for all nine electrodes (as a region of interest) for each condition showing the differences in both amplitude and latency (see dashed box) between HF words, LF words, and PWs. Box‐and‐whisker plots showing the distribution of P600 amplitudes (C) and latency (D) across language conditions (HF words, LF words, and PWs). For each language condition, separate box plots are shown for fast readers (black) and slow readers (red). The box plots illustrate the median, mean, interquartile range (IQR), and range of the amplitude data. Outliers are not explicitly marked but are contained within the range.