Baby Mind-Reading and False Visual Memories


Introducing the Brain Byte: bringing you something to think about in 5 minutes or less. For this first installment, we draw inspiration from the BPS Research Digest–check it out! Our podcast transcript is below.

Turns out he's really listening.

Babies can tell whether you’ve made a mistake from the tone of your voice. This is a study by Elena Sakkalou and Merideth Gattis at Cardiff University, published in the journal Cognitive Development. The study had two parts. The first part replicated an earlier study. Twenty-eight toddlers aged 14 to 18-months saw an experimenter perform two actions on a toy (for example, pushing it or rolling it), one of which was accompanied by the word “There” as if the action were intended; the other by “Whoops”. The experimenter always kept a neutral facial expression. Afterwards the infants could handle the toy themselves, and they were more likely to imitate the action that was accompanied by the word “There” It’s as if they knew the adult intended that action.

The second part of the study had the same setup except that two Greek words were used and 56 infants participated. Crucially, they were spoken with the same prosody—that is, the same profile of volume, duration and pitch. None of these toddlers came from Greek-speaking homes. It turned out the toddlers older than 16 months still imitated more “intentional” actions than accidental actions on the toys. This is a strong indication that they were listening to the way the words were said.

The bottom line? Babies pay attention to cues about people’s intentions and desires, and they pick up on prosody cues around 16 months.

Sakkalou, E., and Gattis, M. (2012). Infants infer intentions from prosody. Cognitive Development, 27 (1), 1-16 DOI: 10.1016/j.cogdev.2011.08.003

You're not watching as closely as you think.

Your memory of events is distorted within seconds. Memory is a creative process, etched in sand, not stone. Brent Strickland and Frank Keil made a dramatic demonstration of this fact in a study at Yale, published in the journal Cognition.

Fifty-eight university students watched three types of 30-second video clip, each featuring a person kicking, throwing, putting or hitting a ball or shuttlecock. All videos were silent. One type of video ended with the consequences of the athletic action implied in the clip – for example, a football flying off into the distance. Another type lacked that final scene and ended instead with an irrelevant shot, for example of a linesman jogging down the line. The final video type was scrambled, with events unfolding in a jumbled order. Crucially, regardless of the video type, sometimes the moment of contact – for example, the kicker actually striking the ball – was shown and sometimes it wasn’t.

After watching each video clip, the participants were shown a series of stills and asked to say if each one had or hadn’t featured in the video they’d just watched. Here’s the main finding. Participants who watched the video type that climaxed with the ball (or shuttlecock etc) flying off into the distance were prone to saying they’d seen the causal moment of contact in the video, even when that particular image had in fact been missing.

In other words, because seeing the ball fly off implied that the kicker (or other protagonist) had struck the ball, the participants tended to invent a memory for having seen that causal action happen, even when they hadn’t. This memory distortion happened within seconds, sometimes as soon as a second after the relevant part of the video had been seen.

This memory invention didn’t happen for the videos that had an irrelevant ending, or that were scrambled. So memory invention was specifically triggered by observing a consequence that implied an earlier causal action. In this case, the participants appeared to have “filled in” the missing moment of contact from the video, thus creating a causally coherent episode package for their memories.

A second study replicated these memory distortion effects with 58 more participants and with new contexts involving kicking, throwing and bowling.

The researchers said their findings have obvious implications for crime scene witnesses. Imagine a witness sees a man wielding a gun, and imagine seconds later they also see a person nearby falling from a gunshot wound – these new results show how easily the mind of the witness could invent a memory of having seen the moment the trigger was actually pulled.

The bottom line? Our memories are reconstructions, and our brain is willing to make up details that seem to make sense.

Strickland, B., and Keil, F. (2011). Event completion: Event based inferences distort memory in a matter of seconds. Cognition, 121 (3), 409-415 DOI: 10.1016/j.cognition.2011.04.007

Download the podcast here.