Abstract
Dimensionality poses a serious challenge when making predictions from human neuroimaging data. Across imaging modalities, large pools of potential neural features (e.g., responses from particular voxels, electrodes, and temporal windows) have to be related to typically limited sets of stimuli and samples. In recent years, zero-shot prediction models have been introduced for mapping between neural signals and semantic attributes, which allows for classification of stimulus classes not explicitly included in the training set. While choices about feature selection can have a substantial impact when closed-set accuracy, open-set robustness, and runtime are competing design objectives, no systematic study of feature selection for these models has been reported. Instead, a relatively straightforward feature stability approach has been adopted and successfully applied across models and imaging modalities. To characterize the tradeoffs in feature selection for zero-shot learning, we compared correlation-based stability to several other feature selection techniques on comparable data sets from two distinct imaging modalities: functional Magnetic Resonance Imaging and Electrocorticography. While most of the feature selection methods resulted in similar zero-shot prediction accuracies and spatial/spectral patterns of selected features, there was one exception; A novel feature/attribute correlation approach was able to achieve those accuracies with far fewer features, suggesting the potential for simpler prediction models that yield high zero-shot classification accuracy.
Highlights
The curse of dimensionality is severe in neuroimaging data, and prediction algorithms trained on neural data must take this into account to avoid overfitting
The zero-shot problem was simulated by employing LOCO cross-validation; feature selection and training was performed using 59 of the 60 classes, and one class was held out for testing
The efficacy of the feature selection techniques were compared in terms of prediction accuracy as well as in the locations or the frequencies of the features that were selected
Summary
The curse of dimensionality is severe in neuroimaging data, and prediction algorithms trained on neural data must take this into account to avoid overfitting. There are often very large sets of potential neural features or dimensions, and they are often recorded across a relatively limited set of stimuli and samples. In functional magnetic resonance imaging (fMRI), responses from tens of thousands of voxels (or more) are commonly analyzed over multiple time points. Magnetoencephalography (MEG), electroencephalography (EEG), and electrocorticography (ECoG) involve only several hundred channels at most, but when combined with high sampling rates and rapidly varying neural responses, the resulting dimensionality is often similar to fMRI. This imbalance between features and samples is a common burden in hypothesis testing and model estimation in neuroscience.
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
