Researchers led by Maryam Shanechi, holder of the Sawchuk Chair in Electrical and Computer Engineering and founder of the USC Centre for Neurotechnology, have created a novel AI algorithm capable of distinguishing brain patterns linked to specific behaviors. Researchers have disclosed a breakthrough in the field of brain-computer interfaces, potentially leading to enhanced understanding of neural connections and novel cognitive patterns, as outlined in a recent publication.
As you immerse yourself in this narrative, various cognitive processes start to unfold, demanding attention and scrutiny.
As you lift your arm to grasp a steaming cup of espresso, the task at hand momentarily forgotten, you pause to summarize the article aloud for your colleague, while simultaneously acknowledging a growing pang of hunger. Here is the rewritten text:
In your mind, a complex array of behaviors, likened to various arm movements, coexists with diverse linguistic expressions and internal emotional states akin to hunger pangs. This simultaneous encoding gives rise to extremely complex and intricate patterns within the mind’s electrical activity. The challenge lies in distinguishing the cognitive templates associated with habitual actions, akin to manual gestures, from an array of disparate mental patterns.
For instance, this dissociation is crucial for developing brain-computer interfaces aimed at restoring motor function in individuals with paralysis. When individuals are captivated by making a motion, they often struggle to articulate their thoughts to their muscles. To restore function in individuals with motor disorders, brain-computer interfaces instantaneously decipher deliberate motions from mental exercises, translating them into commands for external devices, such as robotic arms or computer cursors.
Shanechi and her former Ph.D. Scholar Omid Sani, currently an analysis affiliate in his laboratory, created a novel AI algorithm to overcome this challenge. The DPAD algorithm, dubbed “Dissociative Prioritized Assessment of Dynamics,” effectively captures the essence of its innovative methodology.
DPAD’s artificial intelligence algorithm effectively disentangles the neural patterns associated with curiosity-driven behaviors, isolating them from concurrent mental processes, explains Shanechi. This enables more accurate decoding of mental actions from brain exercises, potentially enhancing the performance of brain-computer interfaces. Additionally, our approach has the capacity to reveal previously unknown mental patterns that would otherwise remain undetected.
“A crucial component of the AI algorithm is the identification of mental patterns linked to curiosity, which are then prioritized during training of a deep neural network, as revealed by Sani.” Once this step is completed, the algorithm can then be trained on any remaining patterns to ensure they do not obscure or interfere with the behavior-related patterns. Neural networks afford remarkable adaptability in modeling various cognitive patterns.
With its versatility, this algorithm holds significant potential for future applications in decoding psychological states such as pain or depression, potentially offering a profound insight into human emotional experiences. By tracking an individual’s symptoms in real-time, personalized therapy approaches can be developed to effectively address their unique needs and improve overall mental wellness outcomes.
“We’re thrilled to pioneer new applications of our methodology that could track symptom progression in mental wellness scenarios,” Shanechi said. Can non-invasive brain-computer interfaces potentially revolutionize the treatment of psychological wellbeing issues, offering a new dimension to neurorehabilitation?
