SURF

S. Zayd Enam

Using Machine Learning Techniques to Fit Receptive Fields of Speech Spectrogram Trained Auditory Neurons

Using the tools of machine learning one can determine parameters of a model that probabilistically best fit experimentally collected data. This gives us insight into determining the model that best describes the data. I will be using machine learning to fit models of receptive fields of auditory neurons (a receptive field is any stimulus that maximizes activation of a neuron) that have been generated using a sparse-coding model. In this case, a sparse-coding model is one that minimizes the number of neurons required to represent different sounds. Sparse-coding models have been shown to accurately predict the receptive fields of auditory neurons in the Inferior Colliculus. It will be my job to determine a mathematical model that best describes the receptive fields of these neurons.

Message to Sponsor

I am extremely thankful to the SURF Rose Hills committee for awarding this fellowship. It has granted me the freedom to focus on a neuroscience research problem that I am passionate about and has given me the resources to tackle this problem with a powerful support structure. This award will do wonders to the refinement of my research abilities and will help me focus on the long-term goal of determining the big problems I will tackle in my research career. Thank you for making working on world-class research accessible to undergraduates.
  • Major: Electrical Engineering, Computer Science
  • Sponsor: Rose Hills Foundation
  • Mentor: Michael DeWeese, Electrical Engineering & Computer Science