How Non-Stop Learning Aids Our Decision Making

Human decision making is seldom consistent, and it’s common for us to make different choices in different circumstances.  In neurological terms, this variability is often attributed to noise, that influences how our brain processes information.  A new study from Harvard Medical School suggests things might not be so straightforward, and this variability might actually be a previously overlooked behavioral strategy.

“It all started with a simple experiment,” the researchers say. “We took two different odours and created several mixtures of the two. During the experiment, the different mixtures were presented to the rats, one at a time. On each trial, the rats had to report which of two odours was more dominant. If it thought the answer was odour A, it would approach a water spout on the right, and if it opted for odour B, it would go to the left. Some mixtures had much more of one odour compared with the other, making it easier to tell which was more salient. Whereas in other mixtures, the difference was more subtle. If the rat got the correct answer, it received a water reward.”

The speed of response of the rats were recorded, along with whether their answer was correct.  The data revealed that the decision making processes didn’t appear to have much in common.  For instance, commonly, harder decisions require longer deliberation, but in the experiment, the rats would take the same amount of time regardless of the complexity of the decision.

“The explanation for this unexpected observation wasn’t easy to come by,” the researchers say. “Finally, we found it by constructing a mathematical model that united separate branches in the field of decision-making. In a sense, our goal was to replicate the rats’ behaviour in a ‘machine’s brain’ with the hope of discovering the underlying variables that produced this surprising result.”

Unexpected strategies

The data revealed an unusual and unexpected strategy, as the rats seemed to adjust their behavior according to the results of previous trials.  In other words, it was learning from past outcomes.

The researchers explain that this is consistent with our continuously changing world, which forces animals to constantly update their decision-making processes on a case-by-case basis.  It’s a behavior that can appear variable from the outside, but is actually a rational case of adapting to our circumstances.

“There isn’t just one type of sensory discrimination task,” the researchers explain. “Various elements in the design of the task may draw out different decision-making strategies. For instance, if we had asked rats to localize the side where a sound comes from instead of discriminating between odours, their strategy would have aligned with our initial expectation. This is because there is a ‘built-in’ right-left category in the brain for certain sensory modalities that are naturally spatially separated, but that’s not the case for olfaction.”

Confidence is also likely to play a part, with the rats evaluating their past decisions and adjusting their subsequent behaviors accordingly.  For instance, if we’re confident, and then make a correct decision, we don’t tend to learn much, but if we’re confident and make the wrong call, it provides us with a chance to learn, which is what happened with the rats.

Suffice to say, the experiment was done in rats, so it’s far from certain that the crossover will apply to human decision making, but the researchers believe it nonetheless offers an interesting starting point to reassess some of the neurological beliefs we have about decision making.

“We believe that our work is a good starting point for exploring further how different subfields of decision-making may interact. We also hope that other scientists will use and refine our models in follow-up experiments. It would be fascinating and informative to see when, how and why our model starts to fail. Making an error is an opportunity for learning something new, and that is both the result and take-home message of our study,” they conclude.

Facebooktwitterredditpinterestlinkedinmail