The neuronal substrates of human olfactory based kin recognition

Abstract

Kin recognition, an evolutionary phenomenon ubiquitous among phyla, is thought to promote an individual's genes by facilitating nepotism and avoidance of inbreeding. Whereas isolating and studying kin recognition mechanisms in humans using auditory and visual stimuli is problematic because of the high degree of conscious recognition of the individual involved, kin recognition based on body odors is done predominantly without conscious recognition. Using this, we mapped the neural substrates of human kin recognition by acquiring measures of regional cerebral blood flow from women smelling the body odors of either their sister or their same-sex friend. The initial behavioral experiment demonstrated that accurate identification of kin is performed with a low conscious recognition. The subsequent neuroimaging experiment demonstrated that olfactory based kin recognition in women recruited the frontal-temporal junction, the insula, and the dorsomedial prefrontal cortex; the latter area is implicated in the coding of self-referent processing and kin recognition. We further show that the neuronal response is seemingly independent of conscious identification of the individual source, demonstrating that humans have an odor based kin detection system akin to what has been shown for other mammals.

Figure 1

A: Average identification performance in each body odor category in Experiment 1. The dotted line represents chance performance and error bars denote standard error of the mean (SEM). B: Confidence judgments plotted against actual proportion of correct answers for the two body odor categories. The dotted line represents perfect calibration, indicating that subjective experience is well matched to actual performance. Values above the line indicate underconfidence in performance, whereas values below the line indicate overconfidence. Note the relationship between excellent identification performance and large underconfidence when subjects were identifying body odors from sister. C: Group averaged ratings of perceived intensity and pleasantness of the three experimental odors in Experiment 2. Error bars in graph denote standard error of the mean (SEM).

Figure 2

A statistical parametric map (t statistics as represented by the color scale) showing group averaged rCBF response to the processing of the odor control [contrast odor control vs. baseline] superimposed on group averaged anatomical MRI. Significant increase of rCBF is seen bilaterally in the orbitofrontal cortex (x, y, z: −13, 22, −14; t = 3.68 and 23, 34, −8; t = 3.81). Coordinates in figure and figure legend denotes slice and peak activations expressed according to the MNI coordinates system. Left in figure represents left side (L).

Figure 3

Statistical parametric maps (t statistics as represented by the color scale) of group averaged rCBF responses to kin recognition [contrast sister vs. friend] superimposed on group averaged anatomical MRI. Graphs represent extracted baseline‐corrected rCBF values within the activation peak in question, using a 7 mm volume of interest (VOI) search sphere, in each odor category. Error bars represent standard error of the mean (SEM). All statistical parametric maps are thresholded at t = 2.5. A: Increased rCBF in the dorsomedial prefrontal cortex (X = −7). B: Increased rCBF in the superior frontal gyrus (SMA) as marked by the yellow circle (Y = −18). C: Increased rCBF in the frontal‐temporal junction as marked by the yellow circle (Y = −11). Left in figure represents left side.