Reduced Benefit from Long-term Item Frequency Contributes to Short-term Memory Deficits in Dyslexia
Eva Kimel, Luba Daikhin, Hilla Jakoby, Merav Ahissar
Memory & Cognition (2024)
Layman summary:
Dyslexia, a specific learning difficulty in acquiring proficient reading, is also characterized by reduced short-term memory. But short-term memory doesn’t operate in isolation: Long-term knowledge plays a significant role in how efficiently short-term memory is used. In this paper we examined the hypothesis that reduced short-term memory in dyslexia (to a large extent) reflects reduced benefits from long-term exposure to items.
Individuals with dyslexia were shown to benefit less from repeated exposure to sounds and syllables. Since exposure shapes long-term memory representations, these reduced benefits could also reduce their relative performance in short-term memory tasks. This is especially true when frequent items, to which the participants were exposed a lot, such as digit-words, are used. In this study, we recruited adult participants with and without dyslexia. We asked the participants to remember a sequence of digits and then repeat it. Each participant performed the task twice: once in their native language and once in their second language. Since exposure to the native language is greater than exposure to the second language, we predicted that the performance in the task administered in the native language would reveal a larger group difference than that in the second language, due to the reduced benefit of item frequency/exposure in dyslexia.
The prediction was confirmed, supporting our hypothesis that reduced short-term memory in dyslexia reflects reduced benefits from long-term exposure to items, and that group difference gets bigger with time.
Figure 1: An illustration of performance in the Digit Span task. The digit Span task is an example of a task that benefits from exposure to its comprising items: Performance increases with every exposure, and thus it is better in a high-exposure condition compared to a low-exposure condition. Performance of adults with dyslexia is reduced when compared to that of adults without dyslexia, and although their absolute performance increases with exposure, so does their relative difficulty. Adapted with permission from Kimel et al., (2022), made available by the Creative commons Attribution 4.0 International License.
Frequency-Specific Contributions to Perceptual Priors: Testing the Predictive-Coding Hypothesis
Itay Lieder, Aviel Sulem, Merav Ahissar
iScience (2024): 108946.
Layman summary:
Recent sensory events form perceptual priors that express our implicit expectations and bias our immediate percept, affecting both high- and low-level processing stages. Do these priors operate top-down and thereby bias low-level features toward recently experienced objects (predictive-coding hypothesis), or do priors operate bottom-up and thereby bias low-level features toward low-level representations of recently experienced objects (local memory circuits hypothesis)?
To determine which of these mechanisms underlies the contraction bias of low-level features toward recent perceptual priors in auditory perception, we used, in a 2-tone pitch discrimination task, the “missing fundamental illusion”, which enables the dissociation of low-level components from the high-level pitch. Surprisingly, pure tones were not biased toward previously perceived complex tones’ pitch when the fundamental was missing, but rather toward the physically present harmonics. This result reveals the existence of a bottom-up mechanism, where low-level memory components contribute to perceptual priors and operate independently of recent high-level percepts. This contrasts with the predictions of the predictive-coding hypothesis, which postulates that predictions are generated at high perceptual levels and operate top-down on low-level representations of current stimuli.
In brief, we demonstrate in this study the important bottom-up contribution of low-level memory components to perceptual priors. We posit that this contribution may stabilize the perceptual organization and underlie continuity between similar low-level features belonging to different object categories in the auditory modality.
Figure 1: Schematic illustration of the proposed mechanism underlying pitch contraction across timbre categories (complex or pure tones).
Low-level representations of recent stimuli (exemplified by the frequency components f0, f1, f2 of a complex tone) form local priors that operate bottom-up on low-level features of the current stimuli (exemplified by pure tones near f0, or near f1).
Figure 2: Chart of the proposed model that captures both top-down and bottom-up contributions to auditory perceptual priors.
Low-level frequency-specific representations of recent auditory stimuli form priors that operate bottom-up (blue) on low-level features of the current stimuli, independent of high-level predictions. In addition, pitch contraction appears to be larger within than across timbre categories, indicating a top-down contribution (red), driven by object-level priors.