Premium
A new method for automated and unbiased classification of respiratory‐related electromyogram (EMG) and pressure waveforms
Author(s) -
Sunshine Michael Daniel,
Fuller David Dwight
Publication year - 2018
Publication title -
the faseb journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.709
H-Index - 277
eISSN - 1530-6860
pISSN - 0892-6638
DOI - 10.1096/fasebj.2018.32.1_supplement.913.7
Subject(s) - waveform , sniffing , breathing , plethysmograph , amplitude , respiratory rate , diaphragm (acoustics) , computer science , medicine , pattern recognition (psychology) , speech recognition , biomedical engineering , acoustics , anesthesia , artificial intelligence , heart rate , blood pressure , physics , anatomy , quantum mechanics , loudspeaker , telecommunications , radar
Respiratory muscle EMG activity recorded with intramuscular electrodes and pressure waveforms recorded using whole body plethysmography are common outcome measures in pre‐clinical studies of respiratory control in awake‐behaving animals. These methodologies produce complicated waveforms which are dynamically changing on short time scales. Indeed, it is rare for an awake rodent to breathe with the regular and “metronomic” pattern that often typifies breathing in anesthetized rodents and respiratory motor patterns observed in vitro. However, many studies in rodent models ignore the complex behavioral influences on respiratory waveforms, and only limited periods of “stable breathing” are chosen for detailed analyses. In ongoing studies, we are developing an unbiased, automated classification system to compare the features (e.g., amplitude, duration), frequency (number of occurrences), and patterning (presentation order) of respiratory‐related EMG and pressure waveforms. The detection algorithm weights specific features of each waveform including: peak‐to‐peak amplitude, time to peak amplitude, area under the curve, and total duration of the event. K‐means clustering is then used to separate the different waveforms into distinct groups. The algorithm was validated using pressure waveforms collected via whole body plethysmography or intramuscular diaphragm EMG in awake adult male rats. For both EMG and pressure data, the algorithm was able to detect and quantify the occurrence of distinct waveforms likely associated with sniffing and tidal breathing. The waveforms were automatically stratified based on waveform characteristics such as amplitude and rate of change (e.g., incrementing or decrementing patterns). This analysis revealed at least six distinct waveforms for both the pressure and EMG data. We conclude that the method can rapidly stratify and quantitate respiratory‐related waveform data in a completely unbiased manner. Support or Funding Information T32‐HD043730 (MDS). SPARC OT2 OD023854 (DDF). 1 R01 NS080180‐01A1 (DDF) This abstract is from the Experimental Biology 2018 Meeting. There is no full text article associated with this abstract published in The FASEB Journal .