Related Experiment Videos
Classification of neural signals by a generalized correlation classifier based on radial basis functions
Alexander Kremper1, Thomas Schanze, Reinhard Eckhorn
1Neurophysics Group, Physics Department, Philipps-University, Renthof 7, D-35032, Marburg, Germany. alexander.kremper@physik.uni-marburg.de
Journal of Neuroscience Methods
|June 5, 2002
Summary
This study introduces a new dimension reduction technique using radial basis functions (RBF) for neuroscience data. The method simplifies complex, high-dimensional data for better classification of sensory responses.
Area of Science:
- Neuroscience
- Machine Learning
- Data Analysis
Background:
- High-dimensional data in neuroscience presents challenges for classification.
- Existing multivariate analysis techniques can be complex or overlook data structures.
Purpose of the Study:
- To develop an efficient dimension reduction method for neuroscience data.
- To improve the segregation of neural measurements into different classes.
Main Methods:
- Developed a dimension reduction technique using radial basis functions (RBF).
- The method involves solving a system of linear equations.
- The approach extends linear correlation-based classification.
Main Results:
- Demonstrated the technique's validity and reliability on artificial datasets.
- Successfully discriminated between neural recordings from monkey visual cortex based on different stimuli.
Conclusions:
- The RBF-based dimension reduction method offers a computationally efficient solution.
- This technique is effective for analyzing complex neuroscience data and classifying neural responses.