Separating the FN400 and N400 event-related potential components in masked word priming.

Masked word repetition (priming) increases "old" responses on an episodic recognition test, which has been attributed to more fluent target processing. Such results hinge on comparisons to a control prime that is "fluency-neutral". A common practice is to use unrelated word primes for this purpose when some evidence suggests that they actually decrease target word processing fluency (disfluency). ERP and behavioral measures were collected in three experiments that used non-letter symbols as a fluency-neutral control and match primes to increase processing fluency. Experiment 1 compared unrelated word primes and orthographically dissimilar nonword primes to determine whether these primes cause disfluency. Experiment 2 contrasted orthographically dissimilar and similar nonword primes. Experiment 3 examined semantically related primes to test theoretical predictions derived from Experiments 1 and 2. All three experiments provide evidence that the FN400 and N400 are distinct ERP components because many primes altered only one of the components. Relative to the control condition, match (Exps 1 & 2) and semantic primes selectively affected N400 amplitudes, whereas unrelated word primes and orthographically dissimilar nonword primes selectively affected FN400 amplitudes. The Unexpected Fluency Attribution model (Mecklinger & Bader, 2020) provides a framework for understanding the cognitive processes associated with each ERP component.