The Effects of Physical Exercise and Cognitive Training on Memory and Neurotrophic Factors Jennifer J. Heisz, Ilana B. Clark, Katija Bonin, Emily M. Paolucci, Bernadeta Michalski, Suzanna Becker, and Margaret Fahnestock Abstract ■ This study examined the combined effect of physical exer- cise and cognitive training on memory and neurotrophic factors in healthy, young adults. Ninety-five participants completed 6 weeks of exercise training, combined exercise and cognitive training, or no training (control). Both the exercise and com- bined training groups improved performance on a high- interference memory task, whereas the control group did not. In contrast, neither training group improved on general recog- nition performance, suggesting that exercise training selectively increases high-interference memory that may be linked to hip- pocampal function. Individuals who experienced greater fitness improvements from the exercise training (i.e., high responders to exercise) also had greater increases in the serum neuro- trophic factors brain-derived neurotrophic factor and insulin- like growth factor-1. These high responders to exercise also had better high-interference memory performance as a result of the combined exercise and cognitive training compared with exercise alone, suggesting that potential synergistic effects might depend on the availability of neurotrophic factors. These findings are especially important, as memory benefits accrued from a relatively short intervention in high-functioning young adults. ■ INTRODUCTION Physical exercise potently stimulates brain function (Heisz, Gould, & McIntosh, 2015; Heisz, Vandermorris, Wu, McIntosh, & Ryan, 2015; Hillman, Erickson, & Kramer, 2008), and its positive effects may be enhanced when com- bined with cognitive training. In animal models, exercise promotes the proliferation of new neurons within the den- tate gyrus of the hippocampus whereas cognitive training promotes the survival and integration of those new neu- rons within the network (Fabel et al., 2009; Olson, Eadie, Ernst, & Christie, 2006; Gould, Beylin, Tanapat, Reeves, & Shors, 1999), suggesting that these synergistic pathways improve hippocampal function. Therefore, an intervention that combines the two may have a greater impact on hip- pocampal function than exercise alone. This study exam- ined the impact of exercise training versus combined exercise and cognitive training to determine whether there are synergistic effects on memory in humans. Neuro- trophic factors that support the survival and function of hippocampal cells were also examined as a potential mech- anism for the observed memory changes. If exercise promotes hippocampal function, it is impor- tant to determine the specific cognitive benefits. The hip- pocampus plays a pivotal role in the formation and retrieval of memories for complex events and episodes, and its subregions subserve different aspects of memory processing (Olsen, Moses, Riggs, & Ryan, 2012). The cornus ammonis 3 subregion (CA3) is critical for associa- tive memory formation and retrieval, such as object– place and odor–place associations (Gilbert & Kesner, 2003) as well as novel place learning (Stupien, Florian, & Roullet, 2003). In contrast, the dentate gyrus is critical for the finer details of memory and is believed to play an important role in resolving interference between highly similar contexts, such as learning to discriminate nearby spatial locations (Hunsaker & Kesner, 2008). One poten- tial mechanism for interference reduction in the dentate gyrus is sparse coding, which can orthogonalize and dec- orrelate incoming information through a process known as “pattern separation” (Rolls, 2013). However, sparse coding does not necessitate better memory. Instead, the opposite has been observed: behavioral discrimination of similar contexts is predicted by the degree of recruit- ment of overlapping populations of dentate gyrus neurons (Marrone, Adams, & Satvat, 2011), suggesting that less pat- tern separation predicts better memory. Newly generated neurons in the dentate gyrus (4–6 weeks old) may reduce pattern separation. These neurons are hyperexcitable, highly plastic, and recruited preferentially into novel mem- ory traces relative to fully mature dentate gyrus neurons (Marín-Burgin, Mongiat, Pardi, & Schinder, 2012; Schmidt-Hieber, Jonas, & Bischofberger, 2004; Snyder, Kee, & Wojtowicz, 2001; Wang, Scott, & Wojtowicz, 2000). Furthermore, computational models suggest that McMaster University, Hamilton, Ontario, Canada © Massachusetts Institute of Technology Journal of Cognitive Neuroscience X:Y, pp. 1–13 doi:10.1162/jocn_a_01164