Neural substrates of subphonemic variation and lexical competition in spoken word recognition

ABSTRACT

In spoken word recognition, subphonemic variation influences lexical activation, with sounds near a category boundary increasing phonetic competition as well as lexical competition. The current study investigated the interplay of these factors using a visual world task in which participants were instructed to look at a picture of an auditory target (e.g. peacock). Eyetracking data indicated that participants were slowed when a voiced onset competitor (e.g. beaker) was also displayed, and this effect was amplified when acoustic-phonetic competition was increased. Simultaneously-collected fMRI data showed that several brain regions were sensitive to the presence of the onset competitor, including the supramarginal, middle temporal, and inferior frontal gyri, and functional connectivity analyses revealed that the coordinated activity of left frontal regions depends on both acoustic-phonetic and lexical factors. Taken together, results suggest a role for frontal brain structures in resolving lexical competition, particularly as atypical acoustic-phonetic information maps on to the lexicon.

KEYWORDS:

• Lexical competition
• phonetic variation
• voice-onset time
• inferior frontal gyrus
• fMRI

Acknowledgements

The authors thank Julie Markant, SR Support (particularly Dan McEchron, Marcus Johnson and Greg Perryman) and Brown MRF staff (specifically Maz DeMayo, Lynn Fanella, Caitlin Melvin and Michael Worden) for help operating the eye tracker and scanner; Corey Cusimano and Neal Fox for assistance with growth curve analysis; and Theresa Desrochers, Nicholas Hindy and Peter Molfese for consultations on fMRI analysis. We also thank Jeffrey Binder and several anonymous reviewers for helpful feedback on previous versions of this manuscript.

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Sahil Luthra http://orcid.org/0000-0002-3517-2609

Notes

1 Prior to conducting the fMRI experiment, we conducted a behavioural pilot experiment (n = 15) outside of the scanner. This pilot employed an analogous design to the one used in the fMRI experiment. While analyses of the pilot data are not reported here, the results were similar to those observed in the fMRI experiment.

2 These additional models used treatment-coded factors instead of the backward difference coding scheme described above. These follow-up models do not differ in their fit to the data; the only difference lies in what is captured by the beta values. In treatment coding, one level (e.g., the shortened level of Acoustic Modification) is set as a reference level; the beta value for the other factor (e.g., Lexical Competition) then reflects a simple effect within that reference level. (In this example, the Lexical Competition beta terms would reflect the simple effect of Lexical Competition for shortened tokens on the intercept, linear, quadratic and cubic terms.) Constructing models for each level of Acoustic Modification allowed us to examine the effect of Lexical Competition separately for each level.

3 Because of concerns that the effects of Acoustic Modification might be driven by differences in overall stimulus length and not by the VOT manipulation, a control analysis was conducted that also included post-consonant stimulus length as a nuisance regressor. The VOT manipulation only affected the duration of the word-initial consonant, so examining post-consonant stimulus length affords us an orthogonal measure of stimulus length and gives us more confidence that the effects of Acoustic Modification reflect our VOT manipulation and not overall differences in stimulus length. All clusters reported in Table 3 also emerged in this follow-up analysis.

4 To account for individual differences in behaviour on functional activation, a control analysis was conducted that included behavioural effect sizes as a continuous covariate in the group-level fMRI analysis. To estimate each subject’s effect size, we extracted subject-by-condition random effects from the second-order model in the growth curve analysis. In particular, we measured for each subject how much larger their competitor effect was in the shortened condition than in the lengthened condition. (Recall that the difference in the lexical competition effect between the shortened and lengthened conditions was the only significant interaction in the eyetracking analysis, and note also that these conditions differ in phonetic competition but not in goodness of fit.) A region in right superior / transverse temporal gyrus [(61, -13, 6), 219 voxels, F = 18.2] was sensitive to the size of this behavioural covariate. For the fixed effects of interest, the same clusters emerged in this control analysis as in the main analysis, albeit at a slightly reduced voxel-level threshold (p < 0.06; 195 voxels required for a cluster-level alpha of 0.05).

Additional information

Funding

Research was supported by National Institutes of Health (NIH) [grant number: R01 DC013064] to EBM and NIH NIDCD [grant number R01 DC006220] to SEB. SG was supported by the Spanish Ministry of Economy and Competitiveness through the Severo Ochoa Programme for Centres/Units of Excellence in R&D [SEV‐2015‐490]. The contents of this paper reflect the views of the authors and not those of the funding agencies.