Methodological Comparisons of Classification Techniques and Refinement of an Operant go/no-go Procedure

• Author / Creator

  Sahu, Prateek K.

• Bioacoustic analyses have been used for a variety of purposes including classifying
  vocalizations for biodiversity monitoring and understanding mechanisms of cognitive processes.
  A wide range of statistical methods, including various automated methods, have been used to
  successfully classify vocalizations based on species, sex, geography, and individual. Here, I used
  several classification techniques, namely Discriminant Function Analyses (DFAs), Support
  Vector Machines (SVMs), and Artificial Neural Networks (ANNs), for sex-based classification
  of zebra finch (Taeniopygia guttata) distance calls using acoustic features measured from
  spectrograms. We found that all three methods (DFAs, SVM, and ANNs) correctly classified the
  calls to respective sex-based categories with high accuracy between 92% and 96%. Frequency
  modulation of ascending frequency, total duration, and end frequency of the distance call were
  the most predictive features underlying this classification in all our models. My results presented
  here agree with previous results and identified highly convergent acoustic features involved in
  this classification. My comparison of classification methods gives researcher much needed
  information to select an optimal classification method.
  Refining and modifying experimental procedures play a vital role in improving
  methodology while also reducing animal distress. Standardized experimental procedures are
  often modified for obtaining optimal responses. Many automated operant conditioning
  procedures have been developed and modified over time. Here, I asked if an increase in feed
  time duration affects discrimination in an operant go/no-go task. Specifically, we used zebra
  finches’ sexually dimorphic distance calls as acoustic stimuli to test whether there were any
  significant differences in performance on an operant discrimination task requiring zebra finches
  iii
  to classify calls according to the sex of the producer when a key experimental parameter, feed
  time duration, was increased from 1 second to 2 seconds. We found no significant differences in
  learning speed (trials to criterion) between birds that were given 1 sec or 2 sec of food access
  following a correct go response. Our results indicate doubling food access duration did not
  impact the speed of acquisition of distance call discrimination in zebra finches. These findings
  suggest that we can provide twice as much time for zebra finches to access food, potentially
  improving animal welfare, whilst still being able to compare new results with historical results.

• Subjects / Keywords

  • bioacoustics
  • songbirds
  • operant conditioning
  • animal welfare
  • audio classification
  • zebra finch
  • Classification methodologies
  • refinement of experiment

• Graduation date

  Fall 2022

• Type of Item

  Thesis

• Degree

  Master of Science

• DOI

  https://doi.org/10.7939/r3-znvv-as29

• License

  This thesis is made available by the University of Alberta Library with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.