The effect of text-picture consistency on memory integration in college students: Evidence from ERPs

In multimedia learning, individuals often need to integrate textual and visual information. The neural mechanisms underlying how consistency between text and pictures affects memory integration remain unclear.

To examine the effect of text-picture consistency on memory integration in college students and to determine the processing time at which this effect occurs.

Totally 28 Chinese college students (n_female = 20, n_male = 8; M_age = 21.64 ± 2.30) participated in this experiment.

An adapted pictorial demonstration paradigm was employed. Participants sequentially learned two pairs of text-picture stem facts (consistent vs. inconsistent), separated by an interval, followed by a recognition test for stem and integrated facts. Behavioural and event-related potential (ERP) data were recorded simultaneously.

Under the text-picture consistency condition, college students showed significantly better memory integration performance than under the inconsistency condition (58% vs. 48%). EEG results showed that the text-picture inconsistency condition (i.e., the picture and text within a single stem fact were semantically mismatched) elicited a larger P600 wave (500-800 ms) than the text-picture consistency condition, suggesting that text-picture inconsistency primarily induces conflict and integration difficulties during late cognitive processing. No significant differences were observed in the early N400 component.

Text-picture consistency facilitates memory integration primarily during late processing, aligning with cognitive theories of multimedia learning. Educationally, it serves as a key quality indicator for multimedia material design. Instructors should prioritize semantic alignment and assess knowledge integration during evaluation.