Measurement effects on critical scaling in neural systems

Area of Science:

• Neuroscience
• Computational Neuroscience
• Data Analysis

Background:

• Phenomenological renormalization group (pRG) analysis reveals scale-free properties in neural population recordings.
• These scale-free properties suggest universal neural behavior across different recording modalities.
• The impact of experimental and analytical choices on pRG scaling remains unclear.

Purpose of the Study:

• To investigate how recording and analysis choices influence phenomenological renormalization group (pRG) scaling exponents.
• To determine the sensitivity of pRG scaling to temporal resolution, measurement nonlinearities, and deconvolution techniques.
• To differentiate scaling driven by neural dynamics from that introduced by measurement artifacts.

Main Methods:

• Utilized a network model known to exhibit scaling under pRG analysis as a testbed.
• Systematically varied parameters such as temporal binning, measurement nonlinearities, and deconvolution.
• Analyzed the resulting changes in inferred scaling exponents and the quality of scaling for cluster covariance eigenvalues.

Main Results:

• Scaling properties derived from pRG analysis are dependent on choices in temporal binning, measurement nonlinearities, and deconvolution.
• The quality of scaling for cluster covariance eigenvalues is particularly sensitive to measurement effects.
• Scaling exponents shift substantially with these transformations, even with identical underlying neural dynamics.

Conclusions:

• Experimental and analytical choices can significantly alter pRG scaling in neural data.
• A framework is provided to distinguish true neural scaling from measurement-induced scaling.
• Understanding these effects is crucial for accurate interpretation of neural population activity.