Have you ever taken a moment or two to search through your memory bank to find the right word? If so, you may want to take one of Abhilasha Kumar’s quizzes.
The assistant professor of psychology is spearheading a collaborative project bringing together researchers in cognitive science aimed at discovering more about how humans search through memory.
This project is jointly funded by the Perception, Action and Cognition program and the Established Program to Stimulate Competitive Research (EPSCoR) at the National Science Foundation (NSF). The total award* amount is just shy of $560,000—nearly $450,000 going to Bowdoin and the remainder to researchers at Indiana University.
Kumar is now working with Bowdoin’s IT department to devise a series of online mind games and quizzes designed to collect data about the workings of our mental dictionary. “Searching for the right words is something that humans do frequently and often effortlessly. This project is designed to help us understand how we do this, and ultimately understand how concepts are organized and retrieved during everyday tasks,” said Kumar.
Another important goal of this research, she added, is to characterize when and how mental search breaks down, such as when people experience word-finding difficulties due to different cognitive impairments. “This can include people who have grown up deaf, or with little to no hearing,” she explained. “They may have a different way of organizing concepts than people with normal hearing. One hypothesis is that due to the lack of sound-related information early on, concepts are not as connected—for example rat and bat may not be as close to each other—which in turn makes them search differently." The hope is that this work will aid in the development and application of clinical measures to help those with various cognitive difficulties.
“A lot of what we’re working on is similar to the game ‘Scattergories,’ often played by families on holiday!” Typically, participants will be asked to compile lists from memory in a given amount of time. “We’re building on well-founded verbal fluency tasks that many people will already have encountered in clinical and research settings, but we want to make them a bit more fun,” said Kumar.
A common test is to name as many animals as you can in a minute, she explained. “If you're in a clinician's office and you're being asked about whether you have problems with memory or language, this is the kind of task they would give you. Then they would look at the total number of animals you’re able to come up with and maybe study the pattern of those items. So, for example they might see if the items are clustered together in a meaningful way. Maybe you’re naming all pets first, then all the farm animals, and then all the wild animals, and so on.”
These types of tests supply researchers with very meaningful information, said Kumar. However, she added, one of the limitations of is that much of it is based on very small data sets. “Clinical data sets are time-consuming to collect, and even the ones that have been used in research studies to understand search processes are to the order of maybe only 500 people or so.” This, explained Kumar, is a fairly small sample and, obviously, the larger the number of participants, the more information researchers have to work with. “We hope to build a more accurate model of memory search that can examine patterns that people have not been able to explore because of the smaller data sets typically collected from students at universities or clinical populations.”
To try to come up with this much larger data set, one of the goals for Kumar and her colleagues is to reframe the task as a game in which thousands of people can participate and receive a score at the end. “That’s why we wrote up this grant proposal, to collect a really large and diverse data set to help us address some of these finer questions about how people group information and what this tells us about how different we are.”
"We’re building on well-founded verbal fluency tasks that many people will already have encountered in clinical and research settings, but we want to make them a bit more fun... A lot of what we’re working on is similar to the game Scattergories."
Kumar and her fellow researchers, which include a number of students, are working with the Bowdoin IT department to devise a web interface. “We want to create a short ten-minute quiz to crowdsource on social media. In addition to memory search challenges such as the Scattergories-type quiz,” said Kumar, “we also want to gauge creativity through the use of what is called the ‘divergent association task.’” This is a psychological test, she explained, designed specifically to measure a person’s creativity by asking them to name ten words that are as different from each other as possible. “It turns out this is very difficult,” added Kumar.
The goal is to share the data set with researchers, as well as make the resources available to clinicians and researchers so they can also analyze their own data using the new web interface that Kumar and her colleagues are currently fine-tuning. They have already worked on making some of these analytic tools openly available via https://forager.research.bowdoin.edu/. She hopes a pilot game will be available within the next two months,
“Last but not least,” she added, “as well as informing the use and interpretation of clinical measures, we plan to promote participation from populations that are typically underrepresented in STEM generally and in cognitive science specifically.”
* Grant number: 2235362
