Limits...
Assessment of a non-invasive high-throughput classifier for behaviours associated with sleep and wake in mice.

Donohue KD, Medonza DC, Crane ER, O'Hara BF - Biomed Eng Online (2008)

Bottom Line: A single Polyvinylidine Difluoride (PVDF) sensor on the cage floor generates signals from motion resulting in pressure.Performance is analyzed through an experiment comparing results from direct human observation and classification of the different behaviours with an automatic classifier used in conjunction with this system.Examples of sequential classifications (2 second increments) over transition regions between sleep and wake behaviour are also presented to demonstrate robust performance to signal variation and explain performance limitations.

View Article: PubMed Central - HTML - PubMed

Affiliation: Department of Electrical and Computer Engineering, University of Kentucky, Lexington, KY 40506, USA. donohue@engr.uky.edu

ABSTRACT
This work presents a non-invasive high-throughput system for automatically detecting characteristic behaviours in mice over extended periods of time, useful for phenotyping experiments. The system classifies time intervals on the order of 2 to 4 seconds as corresponding to motions consistent with either active wake or inactivity associated with sleep. A single Polyvinylidine Difluoride (PVDF) sensor on the cage floor generates signals from motion resulting in pressure. This paper develops a linear classifier based on robust features extracted from normalized power spectra and autocorrelation functions, as well as novel features from the collapsed average (autocorrelation of complex spectrum), which characterize transient and periodic properties of the signal envelope. Performance is analyzed through an experiment comparing results from direct human observation and classification of the different behaviours with an automatic classifier used in conjunction with this system. Experimental results from over 28.5 hours of data from 4 mice indicate a 94% classification rate relative to the human observations. Examples of sequential classifications (2 second increments) over transition regions between sleep and wake behaviour are also presented to demonstrate robust performance to signal variation and explain performance limitations.

Show MeSH
Sensor signal with corresponding sleep-wake decision statistic computed every 2 seconds indicate by X markers. Time series shows transition between rest and sleep and response of classifier decision statistics.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC2365952&req=5

Figure 7: Sensor signal with corresponding sleep-wake decision statistic computed every 2 seconds indicate by X markers. Time series shows transition between rest and sleep and response of classifier decision statistics.

Mentions: Figure 7 shows a case where an epoch of sleep was interrupted by a subtle motion during sleep as indicated by transient behaviour near the hour 36.979. In this case, the mouse shifted to change its sleep position, with the latter position resulting in better contact with the cage floor. In these cases the mouse typically keeps its eyes closed and the observer will not indicate this as a wake state, resulting in another source of discrepancy. However, as described in [6,18], mice exhibit many of short periods of wake during sleep resulting in a power-law distribution for contiguous sleep periods with a reported mean sleep bout of 5.9 minutes, while another study [18] with different criteria found a mean of 0.9 minutes in multiple different inbred strains of mice. Therefore, while this is a discrepancy between human observation and the classifier, it does not imply necessarily the classifier is incorrect. Further study and different methods will be needed to resolve how this transient should be classified.


Assessment of a non-invasive high-throughput classifier for behaviours associated with sleep and wake in mice.

Donohue KD, Medonza DC, Crane ER, O'Hara BF - Biomed Eng Online (2008)

Sensor signal with corresponding sleep-wake decision statistic computed every 2 seconds indicate by X markers. Time series shows transition between rest and sleep and response of classifier decision statistics.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC2365952&req=5

Figure 7: Sensor signal with corresponding sleep-wake decision statistic computed every 2 seconds indicate by X markers. Time series shows transition between rest and sleep and response of classifier decision statistics.
Mentions: Figure 7 shows a case where an epoch of sleep was interrupted by a subtle motion during sleep as indicated by transient behaviour near the hour 36.979. In this case, the mouse shifted to change its sleep position, with the latter position resulting in better contact with the cage floor. In these cases the mouse typically keeps its eyes closed and the observer will not indicate this as a wake state, resulting in another source of discrepancy. However, as described in [6,18], mice exhibit many of short periods of wake during sleep resulting in a power-law distribution for contiguous sleep periods with a reported mean sleep bout of 5.9 minutes, while another study [18] with different criteria found a mean of 0.9 minutes in multiple different inbred strains of mice. Therefore, while this is a discrepancy between human observation and the classifier, it does not imply necessarily the classifier is incorrect. Further study and different methods will be needed to resolve how this transient should be classified.

Bottom Line: A single Polyvinylidine Difluoride (PVDF) sensor on the cage floor generates signals from motion resulting in pressure.Performance is analyzed through an experiment comparing results from direct human observation and classification of the different behaviours with an automatic classifier used in conjunction with this system.Examples of sequential classifications (2 second increments) over transition regions between sleep and wake behaviour are also presented to demonstrate robust performance to signal variation and explain performance limitations.

View Article: PubMed Central - HTML - PubMed

Affiliation: Department of Electrical and Computer Engineering, University of Kentucky, Lexington, KY 40506, USA. donohue@engr.uky.edu

ABSTRACT
This work presents a non-invasive high-throughput system for automatically detecting characteristic behaviours in mice over extended periods of time, useful for phenotyping experiments. The system classifies time intervals on the order of 2 to 4 seconds as corresponding to motions consistent with either active wake or inactivity associated with sleep. A single Polyvinylidine Difluoride (PVDF) sensor on the cage floor generates signals from motion resulting in pressure. This paper develops a linear classifier based on robust features extracted from normalized power spectra and autocorrelation functions, as well as novel features from the collapsed average (autocorrelation of complex spectrum), which characterize transient and periodic properties of the signal envelope. Performance is analyzed through an experiment comparing results from direct human observation and classification of the different behaviours with an automatic classifier used in conjunction with this system. Experimental results from over 28.5 hours of data from 4 mice indicate a 94% classification rate relative to the human observations. Examples of sequential classifications (2 second increments) over transition regions between sleep and wake behaviour are also presented to demonstrate robust performance to signal variation and explain performance limitations.

Show MeSH