Dr. Maria Geffen
Professor University of Pennsylvania
Postdoctoral Fellow Rockefeller University
PhD Harvard University
For Dr. Maria Geffen, computational neuroscience was a perfect marriage of a longstanding interest in psychology and a rivaling passion for math and computer science. Maria had already dabbled in psychology research during her summers throughout high school, but it wasn’t until she was an undergraduate that she came across this field that seemed tailor-made to her interests and skills. She quickly identified and pursued a fundamental question – what are the computational principles that the brain uses to guide perception? – that she continues to explore to this day. Maria is now a major figure in the fields of computational, systems, and auditory neuroscience and a Professor at the University of Pennsylvania in the departments of Otorhinolaryngology, Neuroscience, and Neurology.
Maria’s introduction to computational neuroscience came by way of Dr. John Hopfield, her professor at Princeton and one of the credited founders of the field. She switched her major from psychology to molecular biology with a biophysics minor and started conducting undergraduate research in his lab. She and Dr. Hopfield were thinking about neuronal responses to whisker movements when Maria made a key observation: not all whiskers were the same length. They began to think about the mechanical properties of these whiskers and how their different lengths must confer distinct resonance frequencies – frequencies at which they would naturally oscillate or vibrate with the greatest intensity. Although they developed computational theories and models for how the brain could use this resonance information, it wasn’t until she started her graduate studies in the biophysics PhD program at Harvard that she had the chance to test their ideas. Dr. Chris Moore – a new Assistant Professor at MIT who also studied the rodent whisker system – was impressed by and greatly interested in Maria’s undergraduate work. Maria worked with Chris and a fellow graduate student to measure neural responses to whiskers when they were externally driven at different resonant frequencies. She observed stark tuning to these frequencies in the somatosensory cortex as well as resonant amplification of gentle whisker touch, which was consistent with the theoretical models she had developed as an undergrad.
Barely into graduate school, Maria had already made a significant splash in the field of computational neuroscience, but she was just getting started. She went on to conduct her PhD thesis research in the lab of Dr. Markus Meister, who studies computational principles of visual processing in the retina. The retina provides many practical advantages for studying early stages of vision because it can be removed, secured in a dish, and recorded from electrophysiologically while presenting visual stimuli. For instance, Maria could simulate saccades – rapid gaze-shifting movements of the eyes – by quickly changing the stimuli she presented to the ex vivo retina. During these “pseudo-saccades”, Maria made the surprising discovery that a population of retinal ganglion cells briefly switched from responding to light decrements to responding to light increments (i.e., from “off-cells” to “on-cells”), allowing them to more readily detect shifts in visual scenes.
Having already dabbled in somatosensation and vision, Maria next transitioned to studying audition as a Fellow at the Center for Physics and Biology at Rockefeller University. Rather than working in a single laboratory as a traditional postdoc, this position allowed her to collaborate widely to pursue her own scientific ideas. Because of her interest in auditory processing, she worked primarily with Dr. Marcelo Magnasco to conduct psychophysical studies and develop computational models of perceptual properties of natural sounds. For example, she identified various features that can make a sound like running water sound “natural” versus “unnatural”.
With such extensive and wide-ranging expertise in sensory systems and computational neuroscience, Maria was ready and eager to start her own lab studying computational principles of auditory coding at the University of Pennsylvania. A major focus of her lab is predictive processing in audition and testing the theory of efficient coding: the notion that the brain balances between encoding important stimuli while remaining sensitive to unexpected or behaviorally salient stimuli. In a recent study, they showed for the first time that efficient coding in the auditory cortex facilitates perception and behavior. They tested how well mice could detect sound stimuli with changing levels of background noise while recording from the auditory cortex. Through this work, they discovered that neurons re-scaled their firing according to the different levels of contrast in their auditory environment (imagine a library versus a rock concert) in a way consistent with their theoretical predictions based on a model of efficient coding. One of her lab’s many other research directions focuses on multimodal processing, such as the integration of auditory and visual or somatosensory cues. Maria is particularly enthusiastic about this direction because it allows her to put her diverse background in multiple sensory systems to use.
Although Maria’s scientific training was undeniably marked by numerous and early successes, her journey has not been without struggles. In transitioning from training in well-established labs to starting her own, brand-new lab, she began receiving more negative paper and grant reviews, often with an aggressive or demeaning tone that she had rarely experienced before. This was understandably disheartening and frustrating for Maria, not to mention stressful, since papers and grants are necessary currency for a healthy lab. Moreover, as a woman in the particularly male-dominanted fields of systems and computational neuroscience, Maria has often found herself in the position of being the only woman in the room. It took time for her to learn to make her voice heard, to not be intimidated, and to participate on equal footing in discussions – “that’s not something that was easy for me,” she reflects. She credits making concrete efforts to establish her position in the field, such as by attending and organizing conferences, in helping her make strides on both of these fronts. Moreover, she emphasizes how simply recognizing the biases and imbalances in the academic system can be a powerful shield against internalizing those feelings.
Even in the face of these challenges, Maria has been steadfast in her pursuit of deciphering the neural computations underlying perception and behavior. Since the moment she learned that she could bridge her interests in math and the brain, she hasn’t turned back. “I love the idea that you could write an equation down and then really test whether the brain uses that information in a quantitative way.” Indeed, Maria and her lab are known for writing the equations andtesting them. With innovative work spanning theory and experimentation, Maria has firmly established herself as a leader in her fields.
Find out more about Maria and her lab’s research here.
Listen to Meenakshi’s full interview with Maria on October 23rd, 2023 below!
