Activity-based anorexia is induced through the combination of restricted access to food (typically 1 h per day) and unlimited access to a running wheel. Although these two factors are sufficient to induce the model, a number of methodological issues described below should be considered when designing an experimental paradigm (see Table 1 for an overview).

The activity-based anorexia model stresses the animal; thus, if animals are ordered from a vendor, it is important to allow time to acclimate to the animal facility prior to beginning restricted food access. As a general rule, animals in our lab arrive on postnatal day 21, the onset of adolescence; most animals are not put into experimental conditions until at least 14 days after arrival, although as little as 4 days is sufficient for acclimation in studies when timing is critical, such as when using early adolescent rats. In studies with age as a covariate or independent variable, all animal arrivals should occur on the same postnatal day of age. Animals should be individually housed upon arrival as this provides the most accurate method for measuring individual food intake and wheel running activity (see Note 1). Animals should be handled prior to the start of restricted food access in order to acclimate the animals and investigators to the process and to minimize stress from daily experimental manipulations.

Activity-based anorexia has been induced in multiple strains of rats, i.e., Sprague-Dawley, Long-Evans, and Wistar. To our knowledge, there has not been a systematic study comparing strain-dependent vulnerability to activity-based anorexia in rats. Our lab has utilized Sprague-Dawley rats to induce activity-based anorexia, thus the recommendations described below may be most pertinent to this strain.

Using animals that are the same age and sex helps reduce variability in the behavioral and physiological features of activity-based anorexia. Ideally, animals used in activity-based anorexia experiments should be bred at the institution so that monitoring of early life experiences can occur from the time of birth, in order to identify or control for potential stressors that might influence later development of activity-based anorexia. This is not possible at many institutions, so we recommend ordering animals at an exact day of age with the vendor, identifying specific littermates as part of the randomization procedure included in the experimental design.

Despite the predominately age- and sex-specific onset of anorexia nervosa in adolescent girls, the majority of early studies on activity-based anorexia utilized adult male rats. Existing literature comparing the vulnerability to activity-based anorexia across ages and sex is sparse; both males and females develop activity-based anorexia, but some sex differences may exist in levels of weight loss and hyperactivity (16, 17). The experience in our lab has been that using adolescent female rats provides unique challenges compared to using older rats, given that weight loss and hyperactivity appear to occur much more rapidly and severely during adolescence. To quantify this effect, our lab is currently conducting an NIH-funded study across three different ages in female and male Sprague-Dawley rats to determine the specific vulnerability to activity-based anorexia within this strain of rat. Our preliminary analysis of female data indicates an increased vulnerability to activity-based anorexia when the onset occurs during early and mid-adolescence compared to adulthood, with male data collection still ongoing (Barbarich-Marsteller, unpublished data).

In the traditional sense, the activity-based anorexia model combines unlimited access to a running wheel (24 h per day) with restricted access to food (unlimited access for 1 h per day; see Note 2). Timing of food access (i.e., light vs. dark cycle; at the beginning, middle, or end of the light or dark cycle) does not appear to affect the development of activity-based anorexia across studies, however, separating food access into four 15-min periods over a 24 h period, instead of a single discrete hour, prevents the development of activity-based anorexia in one study, suggesting that discrete food access is a necessary component for the development of activity-based anorexia (18).