Short Description

The human auditory system is a complex network that transforms sound waves into neural signals, enabling the perception and interpretation of auditory information. Understanding this intricate process is a central challenge in computational neuroscience. This thesis aims to develop deep neural network (DNN) models that replicate the brain’s processing of auditory signals at the neuronal level. Drawing inspiration from recent studies by Saddler & McDermott and Kell et al., the project will involve training DNNs on ecologically relevant auditory tasks and evaluating their alignment with behavioral data. The work combines principles from cochlear mechanics, signal processing, and machine learning to gain insights into the computational basis of human hearing.

Your Tasks

• Get familiar with the human auditory system
• Train a DNN to interpret output from an existing peripheral auditory model
• Compare the results to measurement results from human listeners
• provide an extensive documentation of your work

Your Profile

• Motivation and interest in the topic
• Background in Python

Contact

Sebastian Handel (sebastian.handel@tugraz.at or 0316/873 4338) Martin Hagmüller (hagmueller@tugraz.at or 0316/873 4377)