Abstract # 34:

Scheduled for Thursday, June 17, 2010 07:00 PM-09:00 PM: Session 11 (Medallion Ballroom C/D/E/F) Poster Presentation


THE USE OF HIERARCHICAL LINEAR MODELING TO COMPARE LIFESPAN TRAJECTORIES OF BEHAVIOR IN INDIVIDUALLY HOUSED MALE RHESUS MONKEYS WITH AND WITHOUT SELF-INJURIOUS BEHAVIOR (MACACA MULATTA)

B. J. Kelly1, K. M. Stonemetz1, C. A. Major1, A. G. Sayer2 and M. A. Novak1,2
1Harvard Medical School, NEPRC, One Pine Hill Drive, Southborough, MA 01772, USA, 2Psychology Department, Tobin Hall, University of Massachusetts, Amherst, MA 01003, USA
line
     

Previous studies have focused on age-related changes in behavior in socially housed monkeys; however, little is known about how behavior changes across the lifespan in individually housed animals. We collected longitudinal behavioral data on individually housed male rhesus monkeys [N=43] from 2-26 years of age and used Hierarchical Linear Modeling to generate average developmental trajectories for several behavioral categories. Additionally, we explored whether changes in behavior followed different trajectories for monkeys with self-injurious behavior (SIB) than for non-injurious controls. Data were centered at a common age of 10 years. Overall, changes in communicative behavior and general activity followed positive quadratic trends and each showed negative linear slopes at the centered age [communication: slope=-0.096; p<0.001; activity: slope=-0.255; p<0.025]. Interestingly, anxiety-related behaviors followed different trends in SIB and control monkeys. At the intercept, SIB monkeys had higher frequencies of anxiety related behaviors than controls [8.573 vs. 6.705; p<0.05]. The two groups did not differ in their linear slopes at the centered age; however, they followed opposite quadratic trends [SIB: quadratic coefficient=-0.009; Control: quadratic coefficient=0.020; p<0.05]. These data support a hypothesis for differential levels of anxiety in SIB vs. control monkeys persisting across the lifespan. For all models, variance components were significantly different from 0, indicating a large amount of individual variation for each monkey around the expected curve. Supported by NCRR grants #RR011122 and #RR00168.