Why two minds might not be better than one after all

  • Date

    Tue 26 Sep 17

Two minds really aren’t better than one when it comes to making the right decision where there is uncertainty, according to new research.


In fact, the best way to make the right decision is having groups of people deciding in isolation while monitoring their brain activity to tap into their unconscious mind. That way, it can be established who should be trusted more on each decision, leading to more accurate group decisions.

So what did the research uncover?

Lead researcher Dr Davide Valeriani said these findings could lead to some revolutionary new ways of making decisions in the future.

“Our key finding was that two minds aren’t better than one when it comes to a decision where there is a high level of uncertainty and users are able to exchange information during the decision-making process,” he explained. “We found that having people working together as opposed to alone doing the same task does not improve a person’s accuracy: it actually makes it worse. However, tapping into the unconscious mind with BCI (brain computer interfaces) makes communicating and non-communicating groups more accurate than traditional groups.”

So what does this mean for a workplace in the future?

“This could help a variety of workplaces where people have to make decisions under uncertainty,” added Dr Valeriani. “For example, in the financial market, brokers having to decide whether to buy or sell a particular stock could make a decision alone and have a BCI fitted, assessing their trustworthiness before reaching a final, joint decision. In another application, police officers monitoring security cameras could be assisted by the BCIs to better assess the risk present on a scene.”

"As the BCI captures brain signals it taps into the unconscious mind before any other reasoning comes into play"
Dr Davide Valeriani School of computer science and electronic engineering

What did the research involve?

Published in the journal Scientific Reports, the study found that people working with others during a visual-search task made 50% more errors than people working alone. Working with others also significantly reduced the confidence people had in their own decisions.

The study found that by relying on people’s brainwaves – using BCI – to identify how confident someone is in his/her decision can lead to more accurate group decisions. The BCI can predict the person’s confidence in their decision using electroencephalography (EEG), a non-invasive technique to record brain activity. 

“The BCI showed higher levels of activity when people made correct decisions as opposed to when people made an incorrect choice,” explained Dr Valeriani, from our School of Computer Science and Electronic Engineering. “As the BCI captures brain signals it taps into the unconscious mind before any other reasoning comes into play, including the effects of interaction with other decision makers. This, in turn, allows to better assess how likely the user is to have made the correct choice and, then, deciding how much to trust him/her when making the group decision.”

The study involved three visual task experiments where participants were asked whether they could see a polar bear in a crowded picture flashing on the screen for a quarter of a second. Sometimes people worked alone and group decisions were simulated at a later stage (no interaction), and sometimes they worked in pairs influencing each other.

So how do you get the most accurate decision?

The most accurate decisions were found by joining individual decisions made in isolation according to the person’s behaviour (i.e. the time required by the user to cast a vote) and his/her unconscious mind tapped using the EEG. By using these two measures and machine learning techniques the BCI was able to estimate the confidence of each decision maker and decide who to trust more on each decision, leading to more accurate group decisions than traditional groups based on majority voting. 

The researchers also found that when people were allowed to interact they were not able to correctly assess their decision confidence. Typically, people feeling very confident about a decision are more likely to be correct than people feeling less confident. When people were allowed to interact, however, this was not true: the decision confidence reported by each participant after each decision did not correlate with the correctness of that decision.

Explained Dr Valeriani: “In such a situation, group communication is a double-edge sword. For example, it could lead over-confident (but inaccurate) people to convince less-confident (but correct) people to change their opinions towards the wrong decision.”