P4-350: Insulin modulates hippocampal function in early Alzheimer's disease: Preliminary results from the Main study

Abstract

Background: Defective insulin signaling may be associated with sporadic AD. Numerous studies have demonstrated enhanced performance on hippocampal-based cognitive tasks after the administration of both intravenous and intranasal insulin in AD. Insulin receptors are found in high density within the hippocampus. These observations lead us to hypothesize that insulin directly modulates hippocampal function in AD. Methods: These preliminary results from the Memory and Insulin in Early Alzheimer’s Disease (MAIN) study include 3 control and 3 early AD non-diabetic subjects over the age of 60. Each subject underwent fMRI scanning and cognitive testing twice, after double-blind, counterbalanced administration of 40IU insulin aspart and equal volume of saline placebo via nasal spray. During functional scanning subjects were presented a series of pseudorandom indoor/outdoor repeated and novel images blocks and responded by button press whenever an outdoor image was displayed. Fixed-effect GLM analysis (BrainVoyagerQX) was used to compare insulin related changes in BOLD response between the groups. A battery of standard cognitive tests was performed immediately after fMRI scanning. The Wilcoxon signed rank test was used to analyze changes in cognitive performance between the saline and insulin conditions. Results: The fMRI analysis showed greater hippocampal BOLD activity in the insulin condition compared to the saline condition in the AD subjects, while no difference in hippocampal BOLD activity was evident in the control group. The small sample size limits the interpretation of the cognitive testing results, however, analysis of the overall group showed slight mean performance increases on the Logical Memory II test (0.83, p 0.45), and the Selective Reminding Test (trial 1; 0.68, p 0.363, trial 3; 0.67, p 0.712). Conclusions: Insulin appears to differentially increase hippocampal activity in early AD compared to nondemented controls during a memory encoding fMRI stimulus suggesting differences in insulin activity between the groups. The results of this pilot study justify a continued study with a larger sample to better delineate the neural correlates of hyperinsulinemic memory facilitation.