A key issue in language processing is how we recognize and understand words in sentences. Research on sentence reading indicates that the time we need to read a word depends on the amount of (un)expected information that it conveys. On the other side, research on single word recognition shows that each word also has its own recognition dynamics based on the relation between its orthographic form and its meaning. It is not clear, however, how these sentence-level and word-level dynamics interact.

Humans capitalize on statistical cues to discriminate fundamental units of information within complex streams of sensory input. We sought neural evidence for this phenomenon by combining fast periodic visual stimulation (FPVS) and EEG recordings. Skilled readers were exposed to sequences of linguistic items with decreasing familiarity, presented at a fast rate and periodically interleaved with oddballs. Crucially, each sequence comprised stimuli of the same category, and the only distinction between base and oddball items was the frequency of occurrence of individual tokens within a stream. Frequency-domain analyses revealed robust neural responses at the oddball presentation rate in all conditions, reflecting the discrimination between two locally-emerged groups of items purely informed by token frequency. Results provide evidence for a fundamental frequency-tuned mechanism that operates under high temporal constraints and could underpin category bootstrapping. Concurrently, they showcase the potential of FPVS for providing online neural markers of implicit statistical learning.

A pre-print and all data and scripts related to this project are available here.

This paper investigates inflectional suffix priming in Slovene, and compare it to stem priming. Read more in the paper pre-print here.