What do we know:
We can group and split up sounds (and other things) into different categories, which allows us to identify and understand what is in the environment. However, choosing the best category for a sound can be difficult because sounds can belong to more than one category. For example, both dogs and wolves can howl, and incorrectly categorizing a wolf’s howl as a dog’s can be dangerous.
What don’t we know: How we use past experience with similar sounds to categorize new, unfamiliar sounds.
What this study shows: We use previous experience with similar sounds to make a best guess on how to categorize new sounds. Our brain isn’t perfect, though—it has a lot of activity that isn’t related to what we’re experiencing (this is called noise), so our performance isn’t perfect. But it is as good as it can be given the noise in the brain.
What we can do in the future because of this study: Record from neurons in different parts of the brain to determine where experience with previous sounds and their categories are stored, and how and where the brain uses that information to choose a category for new sounds.
Why you should care: Understanding how humans use prior experience to categorize new information will allow us to determine what goes wrong when we make categorization errors. This understanding could also be used to develop tools that can allow computers to perform the same kinds of categorizations, which would be useful for automated object or voice recognition.
The ability to group and segregate sounds (or objects) into different categories is an important process that allows us to simplify and understand the environment. However, determining the best category for a sound can be difficult, as some sounds can belong to multiple categories. For example, both dogs and wolves can howl, and incorrectly categorizing a wolf’s howl as a dog’s can be dangerous. To solve this problem, the brain must use information learned from experience in categorizing similar sounds in the past. We expected that humans would use previous experience to decide on the best category for a new sound, minimizing the chance that they chose wrong. However, we found that humans did not seem to use the best decision strategy that minimized categorization errors. But if we assume that there is noise in the brain that limits the ability to accurately keep track of previous experience, humans’ category choices can be consistent with the best decision strategy.
Categorization is an important process that allows us to simplify, extract meaning from, and respond to sounds (or other objects) in the environment. However, categorization is complicated because a sound can belong to multiple categories. Thus, to choose the best category for a new sound, we must make use of prior information on the categories of similar sounds. Given the importance of categorization, we hypothesized that humans utilize the best decision strategy for making categorical judgments that allows us to minimize categorization errors. However, we found that humans did not minimize errors in their categorization behavior, similar to behaviors exhibited in other perceptual and cognitive tasks. We then explored the bases for this sub-optimal behavior and found that it can be consistent with the best strategy if we assume that humans have trial-by-trial noise in components of the judgment process.