Changes in grooming (G) are not only often seen in behavioral experiments but in some cases are the only parameters that change significantly. Considering that in rodents and some other animals G is activated both is stress and comfort, we have designed an algorithm to distinguish between the effects on G produced by such opposite factors. The first stage of our G-analysis algorithm (GAA) deals with identification and separate registration of certain patterns of a grooming ritual, namely: paw licking (1), nose wash (2), face wash (3), head wash (4), body wash and fur licking (5), leg licking (6), tail and genitals licking and wash (7).

Comfort G is characterized by a quiet progressive transition from one stage to another, which starts from stage 1 and ends up at stage 7. The more stress the more animals demonstrate incomplete (abortive) G acts which do not reach stages 5-7. Time spent at G is also a useful index: the ratio incomplete/completed G (or time spent incomplete G/total time spent at G) would be higher in stressed animals (index [1a,b]).

Further step in G analysis is studying transitions between its stages. In comfort the transition patterns in G are progressive (1-2, 2-3, 3-4, etc.) while hyperactivated stress-related G' stages are rather chaotic and often seen as invalid, i.e. omitted or reversed (3-2, 1-4, 3-1, etc.). The number (%) of "invalid" transitions per total transitions between G stages registered can serve as an additional index [2] of stress or comfort (it is higher in stress).

Stopping of G (interruptions) is also an interesting parameter. Animals in comfort would interrupt their G less frequently and duration of the interruptions will be shorter. Higher interruption rate will indicate anxious arousal state of the animals (but only for short interruptions up to 10 s, as longer interruptions from 15 s can represent freezing, a fearful stress-related nonspecific behavior). Thus, interruptions of G shall be registered separately depending on its duration. The percentage of extra-short and extra-long G interruptions per total interruptions (index [3] and [4]) can reflect both stressful states but present its different aspects (i.e. anxiety/panic- and fear-related compounds, correspondingly).

A synthesis of the latter ideas will be in analysis of G transition interruptions. "Quiet" animals in comfort would more likely interrupt G within the same stage (e.g. 2-INT-2, 3-INT-3) while in stress they would more often interrupt between

the stages (e.g. 3-INT-4). The rate "between"/"within" or "between"/ total transition interruptions (index [5a,b]) will be an indicator of stressful/comfort G.

If past-interruption G stage is reversed (i.e. G continued from stage previous to the one it has started at), the stress state is more likely to consider. The number of reversals after interruptions per total number of interruptions might be a sensitive index [6] for a proper G analysis (it is high for stressed and low for comfort animals).

Finally, in accord to Lorentz's motivational analysis, one might wish to consider non-grooming behaviors (NG) that follow G. They can be split in two groops. If it is freezing, aggression or self-aggression, risk assessment or flight, the stress is likely. For other post-G patterns (exploration, appetitive behaviors, etc) the comfort state is playing a role. Ratio [stress-/comfort-related] post-G NG or total number of stress-related post-G NG per total post-G [G + NG] patterns are additional indicators [7a,b] for interpreting changes of G seen in experiments.

The GAA outcome, spider diagrams with the 7 above indices at 7 different axes, will give a picture of the real nature behind changes in G seen in experimental animals vs. controls. Briefly, the more stress, the bigger square will be at the diagram.

We believe that such an approach will assist in correct interpretation of animal G behavior both in stress and comfort. Moreover, with certain adaptations, it might be used to distinguish between normal, anxiety- and depression-related G in rodents.

© 2000 Noldus Information Technology b.v.