Stress and Memory Bias Interact to Predict Depression in Multiple Sclerosis Joe Beeney and Peter A. Arnett Pennsylvania State University This study is an investigation of the moderating effect of cognitive schema on the relationship between stress and depression in individuals with multiple sclerosis (MS). In the study, the authors employed a performance-based measure of affective memory bias and a self-report measure of everyday stress to assess both direct and interactive effects of cognitive schema and stress on depression in individuals with MS. The specific hypotheses were that high stress would be more highly associated with depression if an individual also demonstrated a bias for negative information, but that a bias for positive information may buffer against the effects of stress on depression. Results supported the hypotheses, demonstrating a significant effect of the interaction and differential effects of stress based on the direction of memory bias. Implications for understanding depression in MS are discussed, as well as dominant theories of adult depression in the general population. The results are also discussed as a potential contradiction to A. T. Beck’s (1967, 1976) developmental hypothesis of cognitive schemas. Keywords: emotional content, human information storage, major depression, stress, memory In the normal brain, myelin sheaths insulate axonal fibers, enabling fluid electrical conduction along the axon between gaps in the myelin. In the brains of those with multiple sclerosis (MS), this process is disrupted (Arnett, 2003; Brassington & Marsh, 1998; Compston et al., 2005). Discrete plaques form, partly caused by proliferating astrocytes, resulting in the destruction, swelling, or fragmentation of myelin. Such plaques can form anywhere in the brain and spinal cord, interrupting saltatory conduction and result- ing in highly variable physical and cognitive impairments. Al- though about three fourths of plaques are found in the white matter, some affect the gray matter and junctions between gray and white matter (Pittock & Lucchinetti, 2007). Common symptoms of this process include muscle weakness, visual disturbances, urinary disturbance, balance problems, and significant problems in any domain of cognition (Arnett, 2003). Depression is much more common among individuals with MS than the general population (Joffe, Lippert, Gray, Sawa, & Hor- vath, 1987; Minden & Schiffer, 1990; Sadovnick et al., 1996), and research suggests depression is more common in individuals with MS than other chronic illnesses, including other neurological dis- orders (Minden, Orav, & Reich, 1987; Schiffer & Babigan, 1984). Before onset of the disease, individuals are no more at risk for depression than the general population. However, after onset of the disease, risk for depression is highly elevated, with lifetime prev- alence estimated at around 50% (Joffe et al., 1987; Siegert & Abernethy, 2005). Understanding what changes are related to this inflation in the prevalence of depression is essential for under- standing the disorder and targeting possibilities for treatment. In the current study, we attempt to assess the importance of the moderating effect of memory bias on the relationship between stress and depressive symptoms in MS. A central goal is to further limited research on the importance of these variables, essential to numerous cognitive models of depression for the general popula- tion, within an MS sample. Various cognitive models have at- tempted to explain the onset and maintenance of depression in the general population (e.g., Abramson, Metalsky, & Alloy, 1989; Abramson, Seligman, & Teasdale, 1978; Beck, 1967, 1976). Vul- nerability to depression in cognitive models (henceforth called cognitive vulnerability) occurs when an individual possesses de- pressogenic cognitive schemas—structures for “screening, coding, and evaluating the stimuli that impinge on the organism” (Beck, 1967, p. 283). According to Beck (1967), these schemas remain latent until activated by stressful life events. When activated, such representations allow greater access to negative themes, shape expectations, evaluations, and perceptions, and they interact with stimuli to guide attention, memory, and cognition (Segal, 1988; Williams, Watts, MacLeod, & Mathews, 1997). Because schemas are thought to interact with information to guide attention and memory, many researchers have used information processing par- adigms to examine negative information processing biases in those with unipolar depression (e.g., Ingram, Bernet, & McLaughlin, 1994; Teasdale & Dent, 1987; Timbremont & Braet, 2004). These paradigms are performance-based procedures for measuring cog- Joe Beeney and Peter A. Arnett, Department of Psychology, Pennsyl- vania State University. This work was presented (in part) at the 35th Annual Meeting of the International Neuropsychological Society in Portland, Oregon, February 2007. It was also part of a master’s thesis from the Pennsylvania State University granted to Joe Beeney. Special thanks to the many neurologists in Pennsylvania, particularly Brian Ahlstrom, who contributed their time to verifying multiple sclerosis diagnoses and ratings of course for the partic- ipants with multiple sclerosis in the project. We would also like to thank Dawn Polen, Lauren Strober, Jared Bruce, Fiona Barwick, Stephanie Pantalone (Marshall), Molly Riley, Luciano Tristan, Martin Pankiewicz, Jess Clark, Karisa Cortellini, Kate Caddick, Lisa Martin, Alfred Baga- masbad, Jennifer Leer (Cromer), Kristin King, La Riena Ralph, Maya Ramirez, Elizabeth Ranft, Hannah Roggenkamp, Megan Wagner, Pinar Miski, Michelle Olson, Erin Stover, and Andrae Laws for their help with various aspects of the project. Correspondence concerning this article should be addressed to Peter A. Arnett, Department of Psychology, Pennsylvania State University, 522 Moore Building, University Park, PA 16802-3105. E-mail: paa6@psu.edu Neuropsychology Copyright 2008 by the American Psychological Association 2008, Vol. 22, No. 1, 118 –126 0894-4105/08/$12.00 DOI: 10.1037/0894-4105.22.1.118 118