How familiarization and repetition modulate the picture naming network

Abstract

A common strategy to reveal the components of the speech production network is to use psycholinguistic manipulations previously tested in behavioral protocols. This often disregards how implementation aspects that are nonessential for interpreting behavior may affect the neural response. We compared the electrophysiological (EEG) signature of two popular picture naming protocols involving either unfamiliar pictures without repetitions or repeated familiar pictures. We observed significant semantic interference effects in behavior but not in the EEG, contrary to some previous findings. Remarkably, the two protocols elicited clearly distinct EEG responses. These were not due to naming latency differences nor did they reflect a homogeneous modulation of amplitude over the trial time-window. The effect of protocol is attributed to the familiarization induced by the first encounter with the materials. Picture naming processes can be substantially modulated by specific protocol requirements controlled by familiarity and, to a much lesser degree, the repetition of materials.

Keywords: Blocked naming; Electrophysiology; Familiarity; Language production; Lexical access; Priming; Semantic interference; Speech.

Figure 1

Event related potentials and statistical test of the contrast between naming protocols. (A) ERPs for the two protocols. For the sake of clarity, only 32 electrodes from among the 60 that were recorded are presented, with a highlight of two representative electrodes showing significant differences (p < .001). (B) Results of running Student t-tests comparing naming protocols performed for every electrode. The tests were iterated over multiple sub-samples of the data to take into account differences in the size of the data sets across protocols; the scale indicates the proportion of sub-samples for which there was a significant difference (p < .001) in either direction (see Methods for details). Red squares delineate the earliest significant differences, interpreted as N 250/N 400 components. Green squares delineate later significant differences.

Figure 2

sLORETA localization of the electrophysiological activity of the grand average for the sequential (A) and the blocked (B) naming protocols at 220-230 ms (top-left) and 420-430 ms (top-right) post-stimulus. (C) sLORETA statistical difference between conditions (based on the localizations computed for each participant in each protocol) for the same time windows and regions of interest. The differences observed were marginally significant across participants (p=.08 for 220-230ms; p = .07 for 420-430ms). Voxels in red correspond to stronger activity in the sequential naming protocol; voxels in blue correspond to stronger activity in the blocked naming protocol. Within each view, black arrows on the margins indicate the location of the two other views.

Figure 3

Event related potentials and statistical test of the repetition effect within the blocked naming protocol. (A) ERPs contrasted for the first two and the last two repetitions on CP1, and P3 (left hemisphere), and CP2, P4 (right hemisphere). Significant effect of repetition noted. The star denotes a significant effect of repetition. (B) Running Student t-test results for all electrodes comparing the first and the last two repetitions within the blocked naming protocol, represented as in Figure 1.

Figure 4

Semantic contrasts in the sequential naming protocol. (A) Average naming latencies in milliseconds for the first two and the last two ordinal positions. A significant semantic interference effect is observed (see also Figure S2). (B) Event related potentials in the sequential naming protocol, with those from the first two and the last two ordinal positions contrasted, on fronto-central, fronto-temporal and centro-parietal and parieto-occipital electrodes of both hemispheres (FC1, FT7, CP1, PO3 and FC2, FT8, CP2, PO4 respectively). There was no semantic effect (i.e., significant difference between the two ordinal position conditions; see also Figure S3).

Figure 5

Semantic contrasts in the blocked naming protocol. (A) Average naming latencies in the two contexts. A significant semantic interference effect is observed (see also Figure S2). (B) ERPs contrasted for homogeneous and heterogeneous blocks, on electrodes FC1, FT7, CP1, and PO3 (left hemisphere), and FC2, FT8, CP2, PO4 (right hemisphere). There was no effect of block homogeneity.