Variability Measures of Positive Random Variables
Author Information
Author(s): Kostal Lubomir, Lansky Petr, Pokora Ondrej
Primary Institution: Institute of Physiology, Academy of Sciences of the Czech Republic
Hypothesis
Can new information-based measures of statistical dispersion better quantify neuronal firing variability than standard deviation?
Conclusion
The proposed entropy-based and Fisher information-based dispersion measures provide better descriptions of neuronal firing variability than standard deviation.
Supporting Evidence
- The study shows that standard deviation does not effectively quantify randomness in neuronal firing.
- Entropy-based and Fisher information-based measures provide different perspectives on neuronal firing variability.
- Experimental data from rat olfactory neurons supports the proposed measures.
Takeaway
This study looks at how we can measure the randomness of neuron firing times better than before, showing that some new methods work better than the old way.
Methodology
The study compares standard deviation with new measures of dispersion based on entropy and Fisher information using both simulated and experimental data.
Potential Biases
Potential biases in the estimation of dispersion coefficients due to the choice of distribution model.
Limitations
The study's sample sizes were small, which may limit the generalizability of the findings.
Participant Demographics
The study involved recordings from olfactory receptor neurons in freely breathing and tracheotomized rats.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website