Visual Face-preference in the Human Fetus?

By Neuroskeptic | June 11, 2017 8:40 am

Even before we’re born, human beings are sensitive to face-like shapes, according to a paper just published in Current Biology.

British researchers Vincent M. Reid and colleagues of the University of Lancaster used lasers to project a pattern of three red dots onto the abdomen of pregnant women. The lasers were bright enough to be visible from inside the womb. The dots were arranged to be either “face-like”, i.e. with two “eyes” above one “mouth”, or inverted. The inverted condition was a control.

reid_dotsBecause the fetus is thought to have better peripheral than central vision, the stimuli were presented at the edge of their visual field and were made to move laterally over about five seconds. Using 4D ultrasound to record fetal movements, Reid et al. found that a fetus was more likely to turn its head towards the moving “face-like” pattern, compared to the inverted control pattern.

reid_fetusThis is a striking result. Previously, it has been shown that three-dot face-like stimuli engage the attention of newborn babies more than inverted stimuli, but this is the first study to look at prenatal visual preferences.

Reid et al. say that while a sensitivity to face-like patterns may be an innate feature of the human brain, it’s also possible that the preference arises from some kind of prenatal visual experience:

These results indicate that the fetus in the third trimester is more likely to engage with stimuli featuring an upright face-like configuration when contrasted with an inverted configuration. We therefore conclude that postnatal experience is not necessary for the emergence of a preferential visual system for facelike stimuli.

This finding rules out rapid postnatal learning, such as filial imprinting, as a mechanism for this visual proclivity. These mechanisms may be innate, or, possibly, the perceptual bias is triggered by exposure to patterned light in the womb during prenatal visual experiences.

The authors stress that their data doesn’t imply that the fetus can actually see the faces of people outside the womb (!)

It should also be noted that the results of the present study do not imply that the fetus can respond to faces presented externally under everyday circumstances. The behavior that has been demonstrated in the current study derives from the specific conditions of the experiment.

In my view this is an interesting study and a very nice experimental paradigm. I did have one concern when I read the paper, related to the methodology. Is it possible that the laser apparatus was exerting a physical pressure on the womb, in such a way that the fetus might detect? Notably, the experimenter holding the stimulator was not blind to the condition (whereas the researchers rating the ultrasound images for head turns were blinded).

I asked the first author, Prof. Reid for a comment on this. He kindly replied, saying that it’s unlikely that physical pressure could explain the results:

Theoretically it is possible that differential pressure could yield a difference in fetal response between conditions. Here are the reasons why I do not think that differential pressure is a plausible explanation for the results that we have reported. Increased pressure would likely induce more overall movement in the “increased pressure” condition than the “lesser pressure” condition. This would produce a loss of ultrasound image of the face in one condition over the other. If the fetus moves too much, the 4D image of the face moves out of the scanning pathway. We did not observe any difference in data quality between the conditions.

Secondly, it does not make sense to me that differential pressure would manifest as a specific behaviour – of orienting towards the stimuli. Overall fetal movement would be the likely response due to increased pressure. Finally, the amount of pressure to gain a differential fetal response would need to be quite substantial and it would need to be consistently differential between conditions. The only way that this could happen would be if the experimenter explicitly and consciously applied substantial pressure in one condition but not in the other. That is something that didn’t happen in this study. Experimental behaviours, including the pressure and speed of stimulus movement, were kept as consistent as possible in order to be replicable.

CATEGORIZED UNDER: faces, papers, select, Top Posts
  • daniele marinazzo

    thanks for this post, and thanks to the authors for sharing the data (btw with a mistyped column label Upright, Away is reported twice).
    The paper is not in my field, but bar plots are a universally suboptimal way to represent data, as said once more in this excellent paper

    Using the code shared by the authors of the paper above, I redid the figures using violin plots and shift functions to compare the distribution. It appears that the difference are not robust across all the quantiles.

    Anyway, this paper should be easy to replicate. If I might say something, why not using a smiley, which is slightly more (human) face-like than a downwards facing triangle?

  • Gideon Nave

    I’m a bit surprised that there’s no manipulation check in a paper that uses an extremely novel method. As far as I could tell, there’s no “baseline” condition (i.e. measurement of how often the babies turn at random). Is there any evidence that the babies are even responding to
    light by movement?

    • Neuroskeptic

      Good point, there was no condition without stimulation. The difference between the upright and inverted suggests that the fetus did respond (to the upright) but it would have been nice to also have a baseline measure of spontaneous movements.

  • Temp

    Interesting and carefully thought out technique – and yet the link with faces seems a little ‘elastic’ – The work with neonates uses dark ‘eye/mouth’ dots stimuli on a white background (e.g., Johnson & Morton, 1991) – something that fits with the Dakin & Watt (2009) horizontal ‘bar code’ view. Yet this work used the inverse – bright eyes & mouth – not sure how a sensitivity to this stimulus would relate to any built-in mechanism to track the visual characteristics associated with face signals.

  • Uncle Al

    4D ultrasound to record fetal movements” I can think of little else more pro-active than frying a fetal brain with 80 – 100 db of ultrasonic disruption. Repeatedly. Diversity!

  • OWilson

    I’m still confused by the opening, “Even before we’re born, human beings are sensitive to face-like shapes”.

    Human beings, or just tissue? :)

    • Neuroskeptic

      Human beings because these were third trimester fetuses, about 240 days since conception. They had been human beings for… roughly two months.

      • OWilson

        Thank you.

        I never get tired of hearing that! :)



No brain. No gain.

About Neuroskeptic

Neuroskeptic is a British neuroscientist who takes a skeptical look at his own field, and beyond. His blog offers a look at the latest developments in neuroscience, psychiatry and psychology through a critical lens.


See More

@Neuro_Skeptic on Twitter


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Collapse bottom bar