Electrophysiological differences in older and younger adults’ anaphoric but not cataphoric pronoun processing in the absence of age-related behavioural slowdown

The question of whether syntactic processing declines with age remains unresolved. While some studies report slower and less accurate performance in older adults on sentence judgment tasks, including pronoun interpretation, others show preserved syntactic processing. These discrepancies might stem from differing methodologies, focusing either on post-interpretive judgments (offline processing) or online sentence comprehension. The 'syntactic decline' versus 'syntactic resilience' debate highlights the need to examine the precise online time course of syntactic processing. Previous research using word-by-word listening tasks suggested preserved online processing in older adults despite poor offline performance. However, these methods do not capture the temporal dynamics of brain activity. This study uses an ERP design to investigate per-millisecond electrophysiological brain activity during anaphoric and cataphoric pronoun processing in young and older French speakers. Pronouns exhibit referential dependency with their antecedents (anaphoric pronouns follow, while cataphoric pronouns precede their antecedents). Gender-mismatch paradigms have revealed disrupted reading times and P600 ERP components for anaphoric pronoun mismatches. Research on cataphoric pronouns is less extensive, with some studies suggesting different processing mechanisms (memory retrieval for anaphors versus antecedent search for cataphors). Prior research on older adults’ pronoun processing indicates potential memory constraints; however, ERP studies are scarce, with mixed findings regarding P600 amplitude differences. The current study aims to clarify the age-related changes in online syntactic processing during pronoun comprehension using ERPs, and addresses the debate by analyzing both behavioral and electrophysiological data, taking into account the potential role of memory constraints.

Existing research on age-related effects on syntactic processing is inconsistent. Some studies report age-related decline in performance on sentence judgment tasks involving pronoun interpretation, potentially due to general cognitive slowdown or specific deficits in working memory-intensive processes. However, other studies demonstrate preserved syntactic processing in older adults, with findings suggesting functional reorganization or recruitment of additional brain regions to compensate for age-related changes. fMRI studies have shown similar activation patterns in young and older adults during sentence processing, though older adults sometimes show increased activity in certain areas. There are also studies demonstrating that older adults perform well on online syntactic processing tasks despite poor offline performance on sentence judgment tasks. The inconsistent findings highlight the need to differentiate between online and offline processing mechanisms and consider the role of individual differences, such as working memory capacity, in shaping age-related performance. The limited research using time-sensitive neurophysiological measurements, specifically ERPs, creates a knowledge gap concerning the precise temporal dynamics of syntactic processing in aging.

Thirty-three native French speakers (18 young adults, mean age 22.05; 15 older adults, mean age 65.73) participated in an ERP study. Participants read sentences with anaphoric or cataphoric pronouns that were either gender-matched or mismatched with their antecedents. A word-by-word moving-window paradigm was used, with participants making acceptability judgments at the end of each sentence. EEG data were recorded using a 64-channel system. Preprocessing steps included downsampling, rereferencing, filtering, artifact rejection using independent component analysis (ICA). Behavioral data (accuracy and response times) were analyzed using mixed-effects regression models. ERP data were analyzed using repeated-measures ANOVAs, focusing on three time windows (300–500 ms, 500–700 ms, and 700–900 ms). Fifteen regions of interest (ROIs) were defined based on standard electrode groupings. Linear regression models examined the relationship between participants' verbal short-term memory (VSTM) spans and P600 amplitudes.

Behaviorally, older adults showed similar accuracy and response times to younger adults, with no significant age-related differences in detecting gender mismatches. Older adults performed slightly better on cataphoric pronoun conditions. Individual differences in accuracy, modulated by VSTM scores, were observed in both groups. ERPs revealed a central/posterior P600 component in both anaphoric and cataphoric conditions, with similar timing and amplitude across age groups. However, older adults showed an additional anterior negativity (500–700 ms) specifically for anaphoric mismatches. The P600 amplitude in the anaphoric mismatch condition was significantly correlated with VSTM scores in older adults, but not in younger adults. This suggests that older adults with higher VSTM capacity exhibited larger P600 amplitudes. No significant group differences were found in the P600 amplitude for cataphoric conditions. For anaphoric pronouns, while both age groups had increased positivity (P600) for mismatches, the effect was more sustained in younger adults.

The findings challenge the notion of a uniform decline in syntactic processing with age. Older adults exhibited comparable behavioral performance to younger adults, suggesting that syntactic processing may be resilient to the effects of aging. The additional anterior negativity observed in older adults during anaphora processing may reflect compensatory recruitment of additional neural resources to manage increased processing demands related to memory retrieval. The correlation between P600 amplitude and VSTM scores in older adults further supports this interpretation. The lack of such a correlation in younger adults suggests that the compensatory mechanism is specific to older adults with potentially lower memory resources. The absence of significant age differences in the P600 amplitude for cataphoric processing suggests that the compensatory mechanism might be specific to the cognitive processes involved in anaphora resolution (retrieving the antecedent from memory), not in cataphoric resolution (searching for the antecedent). This aligns with the idea that anaphoric and cataphoric processes may rely on different cognitive mechanisms. The study also highlights the importance of considering individual differences, particularly VSTM capacity, when investigating age-related changes in language processing. The observed differences in ERP responses between the age groups, despite comparable behavioural performance, underline the value of electrophysiological measures in identifying subtle age-related changes in cognitive processes.

This study demonstrates that while the syntactic processing of anaphoric and cataphoric pronouns is largely preserved in healthy older adults compared to younger adults, there is evidence of compensatory brain activity in older adults, particularly during anaphora processing, possibly due to age-related decline in working memory. Future studies could investigate the neural correlates of this compensatory mechanism, for instance by using fMRI or other neuroimaging methods to identify specific brain regions involved. Further, future research could explore the effects of different types of pronouns or more complex syntactic structures on age-related differences in language processing.

The sample size, especially in the older adult group, is relatively small, limiting the generalizability of the findings. The use of a gender-mismatch paradigm might not fully capture the complexities of real-world pronoun resolution. The study only examined verbal short-term memory, while other cognitive factors could also influence language processing in aging. Furthermore, the lack of detailed assessment of domain-general cognitive abilities beyond memory limits our understanding of the relationship between cognitive decline and the observed compensatory mechanism.