Report: ERPs and localization

Recent work using functional Near InfraRed Spectroscopy has focused a great deal on the emergence of lateralization patterns during processing of sound contrasts. It is often reproached to that field that they ignore the burgeoning literature on event-related potentials (ERP) that preceded it, and which might contain important clues to e.g. lateralization patterns in infants' processing of vowel contrasts. We therefore investigated the literature for this.
    To begin with, most work has studied infants 6 months or younger, who -under most theoretical accounts- have not gained language-specific perception, and thus cannot provide evidence regarding emergent left-dominance for native-like perception. Moreover, even when all studies are included, the evidence for lateralization is weak. As shown in the Table below, only one study documented lateralization supported by a statistical test through a three-way interaction; this one result could even occur by chance given how frequently no significant effect or interaction has been reported. (Certainly, this may be partially explained by the choice of reference sensor, if the asymmetrical signal is also picked up at the ipsilateral mastoid commonly used as referent.) In any case, the EEG field is moving away from sensor-level descriptions of lateralization, as the 'localist' interpretation of signals may not be warranted.

Author Year Group Contrast Factor Significance Reference
Cheour 1995 Newborns y→i and iy yes no ipsi mastoid
Cheour 1996 Preterms y→i and iy yes no ipsi mastoid
Cheour 1997 3mo y→i and iy yes no ipsi mastoid
Cheour 1998 (no new data)

Cheour 1998 6mo and 12mo Finns e→ø and ɤ no

12mo Estonians e→ø and ɤ no

Pihko 1999 Newborns at-risk and control kaa→ka yes no ipsi mastoid
Pihko 1999 6mo at-risk and control
kaa→ka yes yes (1) ipsi mastoid
Leppanen 1999 Newborns at-risk and control kaa→ka no no (2) ipsi mastoid
Friederici 2002 2mo baa↔ba no

Friedrich 2004 2mo baa↔ba yes no avg mastoids
Kujala (MEG) 2004 Newborns a→i NA (3)

Jansson-Verkasalo 2010 6 and 12mo preterm e→ø and ɤ no

6 and 12mo fullterm e→ø and ɤ no

Shafer 2011 Varied ɛ→ɪ NA NA
Shafer 2012 6mo ɛ→ɪ NA NA

(1) Interaction Hemisphere x Stimulus x Group (standard evokes a larger positive response than the deviant; text says "trend" for this to be greater in the left for control and right for at-risk 6-month-olds, but no follow-up ANOVA)
(2) Comment that response is more widely distributed on the left for controls and right for at-risk.
(3) Hemisphere contrast not relevant as measurements were gathered in only one hemisphere in 8 out of the 10 newborns. For the remaining 2, one showed a mismatch in both hemispheres, and the other only on the left.