Does the late positive component reflect successful reading acquisition? A longitudinal ERP study

Christian Wachinger, Susanne Volkmer, Katharina Bublath, Jennifer Bruder, Jürgen Bartling, Gerd Schulte‐KörneNeuroImage 2017 · DOI

TL;DR

This longitudinal study used event-related potentials (ERPs) to investigate the neural basis of reading acquisition in children. They found that dyslexic children show attenuated late positive component (LPC) amplitudes in the left temporo-parietal region during word processing, suggesting an impaired access to phonological representations rather than degraded representations themselves. This difference was evident even before formal reading instruction and highlights the LPC as a potential early marker for reading difficulties.

What Problem Does This Paper Address?

Developmental dyslexia, a common reading disorder, is strongly linked to deficits in phonological processing. However, a key debate in the field is whether these deficits stem from degraded phonological representations (i.e., the internal sound representations are weak or inaccurate) or from impaired access to these representations in the mental lexicon. Previous studies often examined dyslexic individuals who had already acquired some reading skills, making it difficult to disentangle early developmental differences from compensatory strategies or the consequences of years of reading struggle. This paper aims to address this by tracking neural activity (specifically the LPC) longitudinally from pre-literate stages through the early years of reading acquisition, providing a clearer picture of the developmental trajectory and the specific nature of phonological processing deficits in dyslexia.

Methods

The researchers employed a longitudinal event-related potential (ERP) study design, tracking children over five time points from kindergarten to the end of second grade. They used an explicit word processing task where children heard a spoken word and had to determine if it matched a visually presented word. A picture-word matching task served as a control condition.

The core methodological innovation lies in their statistical analysis: instead of averaging ERPs over predefined regions of interest, they used linear mixed-effects models (LME) applied electrode-wise. This "mass-univariate" approach allows for a more precise localization of effects across the entire scalp without a priori assumptions about specific brain regions. They analyzed the difference in the Late Positive Component (LPC), a brain response known to be associated with phonological processing, between the word and picture conditions. The analysis focused on the time window of 600-900 ms, a common period for LPC activity. They specifically investigated group differences (dyslexia vs. control) and the longitudinal development of the LPC.

Datasets

The study utilized data collected from 86 monolingual kindergarten children. Participants were recruited based on familial risk for dyslexia, with the expectation of a higher prevalence of reading difficulties. At the end of second grade, the sample consisted of 64 children. The study did not introduce a new benchmark dataset but rather collected novel longitudinal ERP data from a well-characterized cohort.

Key Insights and Ideas

  • Impaired Access, Not Degraded Representations: The most significant finding is that dyslexic children showed attenuated LPC amplitudes specifically in the word condition, but not in the picture control condition. This suggests that the problem lies not in the quality of their phonological representations themselves, but in their ability to access these representations when processing written words.
  • Early Emergence of Deficits: The differences in LPC amplitude were already present in kindergarten, even before formal reading instruction began, indicating that these neurophysiological differences are not solely a consequence of learning to read.
  • Dynamic Developmental Trajectory: In typically developing children, LPC amplitude increased significantly from kindergarten to first grade (during initial reading acquisition) and then decreased as reading became more automatic. This pattern was absent in dyslexic children, who showed a different, less pronounced developmental trajectory.
  • Left Temporo-Parietal Involvement: The observed differences were primarily localized in the left temporo-parietal region, an area known to be crucial for phonological processing and the grapheme-phoneme mapping involved in reading.
  • LPC as a Potential Early Marker: The findings suggest that LPC attenuation in young children could serve as an early indicator of later reading problems, potentially allowing for earlier intervention.

Key Results

  • Group Differences: Dyslexic children exhibited significantly lower LPC amplitudes in the left temporo-parietal region during word processing compared to control children, particularly in the early stages of reading acquisition.
  • Longitudinal Development (Controls): In control children, LPC amplitude increased from kindergarten (T1) to middle of first grade (T2), then decreased from T2 to the end of second grade (T5).
  • Longitudinal Development (Dyslexia): Dyslexic children showed a different pattern, with a lower initial amplitude and a less pronounced increase/decrease over time compared to controls.
  • Condition Specificity: The group differences were observed in the word processing task but not in the picture processing task, supporting the "impaired access" hypothesis.
  • Early Detection: Significant group differences were already detected at the kindergarten time point (T1).

Limitations and Future Work

  • Sample Size: While the statistical approach was robust, the study had relatively small group sizes (16 dyslexic, 15 control children in the final analysis), which the authors acknowledge as a challenge for detecting subtle effects. However, they note that the significant findings despite this limitation speak to the robustness of the observed differences.
  • Definition of Dyslexia: The study defined groups based on familial risk and later reading fluency, which is a pragmatic approach for longitudinal studies but might not capture all nuances of dyslexia.
  • Future Work: The authors suggest that their findings could inform the development of prevention programs for dyslexic children. They also highlight the potential of the LPC as an early predictive marker for reading difficulties, paving the way for earlier identification and intervention before formal reading instruction begins. Further research could explore the specific neural mechanisms underlying this impaired access and investigate interventions targeting this deficit.

Summary updated

Abstract

Developmental dyslexia is a reading disorder that is associated with deficits in phonological processing, where the exact neural basis for those processing deficits remains unclear. In particular, disagreement exists whether degraded phonological representations or an impaired access to the phonological representations causes these deficits. To investigate this question and to trace changes in neurophysiology during the process of reading acquisition, we designed a longitudinal study with event related potentials (ERPs) in children between kindergarten and second grade. We used an explicit word processing task to elicit the late positive component (LPC), which has been shown to reflect phonological processing. A brain-wide analysis of the LPC with an electrode-wise application of mixed effects models showed significantly attenuated amplitudes in the left temporo-parietal region in dyslexic children. Since these differences were only present in the word and not in the picture (i.e. control) condition, the attenuated amplitudes might reflect impaired access to the phonological representations of words. This was further confirmed by the longitudinal development, which showed a rapid increase in amplitude at the beginning of reading instruction and a decrease with continuing automatization, possibly pointing to a progression from grapheme-phoneme parsing to whole word reading. Our longitudinal study provides the first evidence that it is possible to detect neurophysiological differences in the LPC between children with dyslexia and control children in both preliterate and very early stages of reading acquisition, providing new insights about the neurophysiological development and a potential marker of later reading problems.
BibTeX
@inproceedings{wachinger2017does,
  author = {Christian Wachinger and Susanne Volkmer and Katharina Bublath and Jennifer Bruder and Jürgen Bartling and Gerd Schulte‐Körne},
  title = {{Does the late positive component reflect successful reading acquisition? A longitudinal ERP study}},
  booktitle = {NeuroImage},
  year = {2017},
  doi = {10.1016/j.nicl.2017.10.014},
}
Figure