Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Real-time word anticipation across the lifespan: age-related differences
Poster Session C, Saturday, September 13, 11:00 am - 12:30 pm, Field House
Victoria Cano Sánchez1, Itziar Laka1, Alice Foucart2, Mikel Santesteban1; 1University of the Basque Country (UPV/EHU), 2Nebrija University
Prediction, supported by integration, plays a key role in comprehension (Ferreira & Chantavarin, 2018).This study explores (1) how aging affects lexical cloze-probability during reading, and (2) how language experience (LE),working memory capacity (WMC) and executive control (EC) modulate age-related differences. Predictive processing allows young adults (YAs) to speed up and reduce fixations on high-cloze vs. low-cloze words (Luke & Christianson, 2016). We ask whether this ability declines with age. Existing evidence is mixed: some studies based on the resource-based hypothesis (H1) report that older adults (OAs) predict less efficiently across conditions, showing more fixations and regressions than YAs overall (Schotter, 2012). Others, based on the experience-based hypothesis (H2) argue that although word prediction capacity declines with age, context and individual differences, like LE, result in larger cloze-probability effects for OAs only in low-cloze words, but not in high-cloze ones (Payne et.al., 2021). We conducted an eye-tracking study with 45 YAs (18-34 years; 22.97(3.4)) and 49 OAs (60-80 years; 69.46 (3.7)) healthy native Spanish speakers. Participants read the same Spanish sentences from Foucart et al. (2014) containing target nouns with high- vs. low-cloze probability. WMC was measured via the automated reading span task, EC via the Stroop and Keep Track tasks, and LE with the BEST test and a reading habits questionnaire (RH). Based on prior work (Payne et.al., 2021) we expected that (1) YAs would show faster early integration of high-cloze words (shorter first fixation), while OAs would exhibit larger late-stage cloze-probability effects (more regressions and longer total reading times for low-cloze nouns); and (2) if LE mitigates age-related declines it would be expected to reduce the magnitude of cloze-probability effect mainly for OAs. LMER analyses revealed main cloze-probability effects (larger fixations and more regressions to low-cloze words), WMC (larger fixations with lower WMC) and group (larger fixations and more regressions for OAs than YAs) in the pre-, post- and critical regions. At the critical region, YAs showed a trend for larger cloze-probability effects in early reading measures, whereas OAs exhibited larger effects in later, integrative measures. Cognitive abilities modulated these effects differently across groups: higher WMC enhanced cloze-probability effects in YAs but was associated with reduced effects in OAs. Monitoring & updating capacity influenced reading effort mainly in OAs, with lower rates linked to increased processing difficulty; inhibitory control (stroop) measures did not clearly modulate the cloze-probability effects of the participants. However, LE modulated cloze-probability effects in both groups, with larger modulation in OAs: lower vocabulary and RH were associated with larger cloze-probability effects. The findings support H2 suggesting that (1) cloze-probability is maintained differently across the lifespan and is context-sensitive: YAs predict more efficiently at early stages, while OAs struggle more in integrating low-cloze words later; and (2) cloze-probability effects are modulated by cognitive resources and LE, with WMC and LE playing key roles: lower WMC exacerbate cloze-probability difficulties, while greater LE mitigates those age-related declines.
Topic Areas: Reading, Control, Selection, and Executive Processes