HMN 2025: How some voices are extra memorable than others, no matter who’s listening to them

The time period “memorability” refers back to the probability {that a} explicit stimulus, reminiscent of an object, face or sound, shall be remembered by these uncovered to it. Over bygone days few years, some psychology research have been exploring the extent to which some stimuli are intrinsically extra memorable than others, or in different phrases, whether or not individuals are typically extra more likely to bear in mind them in comparison with different stimuli of the identical sort.

Researchers on the University of Chicago just lately got down to particularly examine the memorability of voices. Their findings, published in Nature Human Behaviour, recommend that some voices are extra memorable than others and their memorability will be persistently predicted throughout totally different listeners.

“Research on intrinsic memorability—the consistencies in what folks bear in mind and neglect—is a reasonably new however lively space of cognitive psychology,” Cambria Revsine, first creator of the paper, advised Medical Xpress. “Many research from our lab and others have extensively explored this phenomenon over bygone days decade, discovering that members have a tendency to recollect the identical pictures of faces, scenes, objects, and way more. However, no prior study to our data has investigated the memorability of auditory stimuli.”

Static pictures and dynamic audio clips have inherent variations and are processed in another way within the mind. Building on their earlier research, by which they discovered that some static pictures of faces, scenes and objects are extra memorable than others, Revsine and her colleagues got down to examine the likelihood that some voices are additionally persistently extra more likely to be remembered or forgotten.

“We additionally aimed to foretell members’ reminiscence efficiency for voices to find out what options make a voice memorable,” defined Revsine. “To check our analysis questions, we used the TIMIT corpus, a large-scale database containing voice clips of a whole bunch of U.S. audio system. In experiments hosted on the net platform Amazon Mechanical Turk, hundreds of members carried out a reminiscence activity on clips of audio system saying the identical sentence.”

The individuals who took half within the staff’s experiment have been requested to hearken to a sequence of audio clips with folks talking, urgent a key on their keyboard every time they acknowledged the voice within the recordings. The researchers then derived a “memorability rating” for every of the voice clips they used of their experiment, by calculating the extent to which members remembered them, on common.

“We then carried out an evaluation that allowed us to find out how comparable random splits of members have been to one another of their reminiscence efficiency, and located vital consistency within the voice clips they remembered and forgot,” mentioned Revsine. “Finally, to foretell this reminiscence efficiency, we measured dozens of acoustic options of the voices, associated to facets like pitch and harmonics, in addition to higher-level options, like dialect and subjectively rated persona traits as measured in a further experiment.”

Revsine and her colleagues subsequently mixed all these acoustic and higher-level options of the voice recordings right into a computational model. They discovered that this model may precisely predict the memorability of voices utilizing numerous options associated to their pitch, loudness, tempo and vowel manufacturing.

“We discovered that individuals are surprisingly comparable within the voices they bear in mind and neglect, no matter totally different past experiences, attentional states, and extra,” mentioned Revsine. “Not solely have been members constant of their reminiscence for voice clips, however they’d constant reminiscence for audio system themselves, throughout a number of spoken sentences. Additionally, our outcomes assist us perceive why we bear in mind sure voices and never others, discovering that sure acoustic properties, like larger pitch and quantity, enhance voice memorability.”

The fascinating outcomes gathered by this analysis staff may quickly contribute to the understanding of how voices are represented within the mind and remembered over time. Moreover, the remark that some voices are extra memorable than others may have numerous sensible implications and might be leveraged in numerous real-world settings.

For instance, it may inform efforts aimed toward assessing or enhancing the reliability of “earwitnesses” throughout trials or might be harnessed to create extra incisive digital assistants, podcasts or audiobooks. In the long run, the staff’s findings may even encourage the event of audio supplies or voice assistants designed to particularly assist sufferers with reminiscence loss.

“Because the present study examined the memorability of voices when audio system mentioned the identical sentences, we are actually investigating what occurs when spoken content material varies—in different phrases, do folks bear in mind what a memorable speaker says higher than a forgettable speaker, no matter what they’re saying?” added Revsine.

“We are additionally excited about exploring the cognitive and neural mechanisms concerned in processing the memorability of a voice. To accomplish that, we plan to make use of pupillometry and neuroimaging measures to check responses to extremely memorable and forgettable audio system.”

© 2025