Table 1.
Cumulative Link Mixed Model (CLMM) on HGS dataset.
Results from CLMM investigating the effects of Pain group, Observer and Horse on Facial Action Units. Adjusted p-values with Benjamini-Hochberg correction in relation to the respective category are reported for each FAUs. Correspondent likelihood ratio test statistics of the CLMM on HGS data are reported in brackets.
Table 2.
Results regarding the inter-observer reliability analysis on HGS.
Average scores of ICC, with the 95%-confidence interval reported in brackets, are presented for each FAU of HGS. Compared to Dalla Costa et al.[19], this analysis was carried out on a different sample of pictures as reported in the methods.
Fig 1.
MCA results reported on the first component for HGS.
The scores (values '0', '1' and '2') of the six FAUs of the HGS and the pain status (“Pain” vs “No Pain”) are presented, as average values, on the first component of the MCA.
Fig 2.
Boxplots of MCA first component scores stratified by pain group and FAUs score.
The boxplots represent the dispersion of the MCA first component scores of the horse pictures, stratified by pain status ("No Pain", "Pain") and by score ("0", "1" and "2") of each FAU.
Fig 3.
Weights of variables in the LDA component for HGS.
Weights for each FAUs of HGS that best discriminate between the “Pain” and “No Pain” groups.
Table 3.
Cumulative Link Mixed Model (CLMM) on MGS dataset.
Results from CLMM investigating the effects of Pain treatment (“Pre-op” VS “Post-Bupivacaine”, “Pre-op” VS “Post-Meloxicam” and “Pre-op” VS “Post-Saline”), Observer and Mouse on Facial Action Units. Adjusted p-values with Benjamini-Hochberg correction in relation to the respective category are reported for each FAUs. Correspondent likelihood ratio test statistics of the CLMM on MGS data are reported in brackets.
Table 4.
Results regarding the inter-observer reliability analysis on MGS.
Average scores of ICC, with the 95%-confidence interval reported in brackets, are presented for each FAU of MGS. This analysis was not carried out in the Leach et al., 2012 study [18].
Fig 4.
MCA results reported on the first component.
The scores (values '0', '1' and '2') of the four FAUs of MGS and the treatment groups (“Pre-op”, “Post-Bupivacaine”, “Post-Meloxicam” and “Post-Saline”) are presented, on average, on the first component of MCA.
Fig 5.
Boxplots of MCA first component scores stratified by treatment group and FAUs score.
The boxplots represent the dispersion of the MCA first component scores of the mouse pictures, stratified by treatment group (“Pre-op”, “Post-Bupivacaine”, “Post-Meloxicam” and “Post-Saline”) and by score ("0", "1" and "2") of each FAU.
Fig 6.
Weights of variables in the LDA component for MGS.
Weight for each FAUs of the MGS that best discriminate between the categories (“no pain status” vs “pain status”).