Associations between iris characteristics and personality in adulthood

There's a body of data, some of it contested, relating eye colour, or eye colour at some ages, with personality traits. Among other things its been reported that there are associations between dark eyes and traits like extroversion (Gentry et al, 1965), and that among kindergarten children below certain ages blue-eyed children were disproportionately likely to be extremely withdrawn (Rubin and Both 1989). Apparently researchers in the area came to agree that eye colour had very little to do with adult personality (e.g. Rubin and Both, 1989).*

There are some speculations regarding mechanisms that might connect eye colour and personality, relating to the fact that production of some substances involved in eye pigmentation is associated with production of other substances (including norepinephrine and cortisol) that are known to have their own associations with behavioural differences.

After reviewing some of the literature just mentioned the authors of this study point out that there are other significant differences between irises besides their colour, including the number and other properties of features called 'Fuchs' crypts', 'contraction furrows' and 'pigment dots' (see figure below). They sought to investigate whether the mechanisms associated with frequency of these properties were also implicated in behaviour. If they were, then there could be systematic correlations between iris features and personality.

Here's a figure illustrating the iris features, from the paper, via MindHacks:

One example of the sort of pleiotropy at issue here is the gene Pax6, which is apparently involved in the process of muscle-cell differentiation in the iris, and hence some of the visible iris features noted above, and also expressed in the brain with one Pax6 mutation being associated with disinhibition, impulsivity and other traits. Pax6 is also associated with production of dopamine neurons, in turn known to be important for a variety of cognitive processes, including some relating to learning, patience, valuation, and response to risk -- all related to personality.

So if some genes are expressed both in ways that bear on what irises look like, and on personality, there might be correlations between the two. This won’t necessarily be the case. There are multiple causal processes on each side, and room for any single factor common to both to get driven away from showing interesting or detectable correlation. There’s an empirical question here, and the present paper has a go at it. There’s a lot of detail that I won’t try to cover – here’s the main story in sketch form.

428 human undergraduates had their irises imaged, and incidence and properties (e.g. size) of the iris features noted above coded. They also completed the NEO PI-R, an instrument based on the ‘Five-Factor’ model of personality (where the five factors are neuroticism, extraversion, openness to experience, agreeableness and conscientiousness). Each factor was assessed in a way that measured six specific factors (in the case of neuroticism the specifics concern: anxiety, angry-hostility, depression, self-consciousness, impulsiveness, vulnerability).

The study included two main analyses. One was a correlational analysis focussed on associations between individual iris characteristics and specific factors in the NEO PI-R. The other was a cluster analysis looking for relationships between overall iris properties and whole personalities. Both found interesting results, even though the effect sizes were (as expected) generally small.

A number of significant correlations were found in the variable based analysis, including between furrows and 'impulsiveness', furrows and 'self-discipline', crypts and 'warmth', crypts and 'positive emotions', and crypts and 'tendermindedness' (I've only listed associations significant at p<0.01 - others were found at p<0.05).

In the person oriented analysis a number of clusters of iris feature types were found, and a number of them had significant correlations with personality measures. It's more difficult to summarise these results in prose, and the pre-publication PDF annoyingly won't allow copying to the clip-board.

This is interesting. There are a few sorts of follow up I'd like to see:

1. The same general thing, but with behavioural (delay discounting) measures of impulsivity.
2. Direct measurement of the genetics, with the same personality measures.
3. Attempt to find out whether people are sensitive (likely unconsciously if at all) to any of this. One could manipulate images of faces replacing or otherwsie changing the irises, and see whether this alone made differences in respect of attractiveness, perceived mate-worthiness, etc. It could be that folks don't notice this, but if there is a decent visible indicator of personality lying around, there's a chance individuals sensitive to it would have been at an advantage.

Here's the text of the abstract:

Variable and person-oriented analyses were used to explore the associations between personality and three previously untested general iris characteristics: crypts, pigment dots and contraction furrows. Personality data, as measured by the NEO PI-R and ratings of iris characteristics from 428 undergraduate students were collected. Crypts were significantly associated with five approach-related behaviors, i.e., feelings, tendermindedness, warmth, trust and positive emotions, whereas furrows were associated with impulsiveness. These findings suggest that because Pax6 induces tissue deficiencies in both the iris and the left anterior cingulate cortex, Pax6 may influence the extent people engage in approach-related behaviors. The results from using a person-oriented analysis suggested that people with different iris configurations tend to develop along different personality trajectories. Future longitudinal studies, twin-studies and genetic association studies, may benefit from collecting iris data and testing candidate genes for crypts and furrows.

This paper has been written up elsewhere, including at MindHacks, and BPS Research Digest.

* The references here are in the first paragraphs of the paper. I'd put them in full here, but as I said above the "article in press" PDF is locked to copying to the clipboard.

LARSSON, M., PEDERSEN, N., STATTIN, H. (2007). Associations between iris characteristics and personality in adulthood. Biological Psychology, 75(2), 165-175. DOI: 10.1016/j.biopsycho.2007.01.007