One of the most impactful findings in computational neuroscience over the past decade is that the object recognition accuracy of deep neural networks (DNNs) correlates with their ability to predict neural responses to natural images in the inferotemporal (IT) cortex. This discovery supported the long-held theory that object recognition is a core objective of the visual cortex, and suggested that more accurate DNNs would serve as better models of IT neuron responses to images. Since then, deep learning has undergone a revolution of scale: billion parameter-scale DNNs trained on billions of images are rivaling or outperforming humans at visual tasks including object recognition. Have today's DNNs become more accurate at predicting IT neuron responses to images as they have grown more accurate at object recognition? Surprisingly, across three independent experiments, we find that this is not the case. DNNs have become progressively worse models of IT as their accuracy has increased on ImageNet. To understand why DNNs experience this trade-off and evaluate if they are still an appropriate paradigm for modeling the visual system, we turn to recordings of IT that capture spatially resolved maps of neuronal activity elicited by natural images. These neuronal activity maps reveal that DNNs trained on ImageNet learn to rely on different visual features than those encoded by IT and that this problem worsens as their accuracy increases. We successfully resolved this issue with the neural harmonizer, a plug-and-play training routine for DNNs that aligns their learned representations with humans. Our results suggest that harmonized DNNs break the trade-off between ImageNet accuracy and neural prediction accuracy that assails current DNNs and offer a path to more accurate models of biological vision. Our work indicates that the standard approach for modeling IT with task-optimized DNNs needs revision, and other biological constraints, including human psychophysics data, are needed to accurately reverse-engineer the visual cortex.

We developed an intelligent web interface that guides users to perform several Parkinson’s disease (PD) motion assessment tests in front of their webcam. After gathering data from 329 participants (N = 199 with PD, N = 130 without PD), we developed a methodology for measuring head motion randomness based on the frequency distribution of the motion. We found PD is associated with significantly higher randomness in side-to-side head motion as measured by the variance and number of large frequency components compared to the age-matched non-PD control group (p = 0.001, d = 0.13). Additionally, in participants taking levodopa (N = 151), the most common drug to treat Parkinson’s, the degree of random side-to-side head motion was found to follow an exponential-decay activity model following the time of the last dose taken (r = −0.404, p = 6e-5). A logistic regression model for classifying PD vs. non-PD groups identified that higher frequency components are more associated with PD. Our findings could potentially be useful toward objectively quantifying differences in head motions that may be due to either PD or PD medications.