Categorical perception of color is lateralized to the right hemisphere in infants, but to the left hemisphere in adults

A. Franklin; G. V. Drivonikou; L. Bevis; I. R. L. Davies; P. Kay; T. Regier

doi:10.1073/pnas.0712286105

New Research In

Contributed by P. Kay, January 3, 2008 (received for review August 27, 2007)

Abstract

Both adults and infants are faster at discriminating between two colors from different categories than two colors from the same category, even when between- and within-category chromatic separation sizes are equated. For adults, this categorical perception (CP) is lateralized; the category effect is stronger for the right visual field (RVF)–left hemisphere (LH) than the left visual field (LVF)–right hemisphere (RH). Converging evidence suggests that the LH bias in color CP in adults is caused by the influence of lexical color codes in the LH. The current study investigates whether prelinguistic color CP is also lateralized to the LH by testing 4- to 6-month-old infants. A colored target was shown on a differently colored background, and time to initiate an eye movement to the target was measured. Target background pairs were either from the same or different categories, but with equal target-background chromatic separations. Infants were faster at initiating an eye movement to targets on different-category than same-category backgrounds, but only for targets in the LVF–RH. In contrast, adults showed a greater category effect when targets were presented to the RVF–LH. These results suggest that whereas color CP is stronger in the LH than RH in adults, prelinguistic CP in infants is lateralized to the RH. The findings suggest that language-driven CP in adults may not build on prelinguistic CP, but that language instead imposes its categories on a LH that is not categorically prepartitioned.

Footnotes

^†To whom correspondence may be addressed: E-mail: a.franklin{at}surrey.ac.uk or paulkay{at}berkeley.edu
Author contributions: A.F., G.V.D., I.R.L.D., P.K., and T.R. designed research; A.F., G.V.D., and L.B. performed research; A.F. and G.V.D. analyzed data; and A.F., I.R.L.D., P.K., and T.R. wrote the paper.
The authors declare no conflict of interest.
↵ ‖ According to the original definition of CP, a requisite was that there is no within-category discrimination at all. However, this definition of CP in current research is no longer used, and CP is now defined as faster or more accurate between-category than within-category discrimination, when between- and within-category separations are equated (1).
↵ ** The measure was normally distributed for both adult data [skew = 0.27, SE skew = 0.54, Z = 0.50 < 1.96; kurtosis = −0.31, SE kurtosis = 1.04, Z = −0.03 <1.96; Shaprio-Wilk (18) = 0.98, P = 0.95] and infant data [skew = 0.48, SE skew = 0.62, Z = 0.78 < 1.96; kurtosis = −0.53, SE kurtosis = 1.19, Z = −0.44 <1.96; Shaprio-Wilk (13) = 0.95, P = 0.62].
↵ †† The picture is complicated by the fact that there are also some instances of RH CP in adults in circumstances that are probably not attributable to trans-callosal transfer. Notably, aphasics exhibit RH but not LH color CP (32), and sometimes normal adults performing visual search with a concurrent verbal interference task similarly exhibit RH CP and no LH CP (31). The relation of these RH CP findings to each other, to RH CP in infants, and to presumably trans-callosally mediated RH CP in adults, remains something of a mystery.
↵ ‡‡ Color category boundaries observed to date in infant CP coincide rather closely with color boundaries arising frequently in the languages of the world, namely green/blue, blue/purple, and purple/pink (8, 9). We suggest that this circumstance is not caused by linguistic categories building on prelinguistic partitions, but instead may be attributable to the same general organizing principles operating independently in both LH and RH, producing similar (but not identical) categorizations of color. Notably, general clustering principles operating over the irregular shape of perceptual color space have been shown to produce color category systems like those found in the world's languages (38) and might in principle also account for prelinguistic CP.
© 2008 by The National Academy of Sciences of the USA