ABSTRACT Obesity is the second leading cause of premature death. Consumption of ultra-processed foods is theorized to be a key cause of obesity. Ultra-processed foods are formulations of cheap industrial sources of dietary energy and nutrients plus additives such as fat, sugar, and flavors that enhance acceptability of the foods. A cross- over experiment with overweight adults found that ad lib access to an ultra-processed diet for 2-weeks resulted in increased caloric intake (508 kcal/day) and more weight gain versus ad lib access to a minimally-processed diet matched for presented calories, energy density, macronutrients, sugar, sodium, and fiber (Hall et al., 2019). The fact that ad lib access to ultra-processed foods resulted in a large increase in caloric intake and weight gain implies that ultra-processed foods may more effectively activate brain regions implicated in reward processing, attention/salience, and memory that influence eating behavior. However, no brain imaging study has experimentally tested whether ultra-processed foods are more effective in activating brain regions implicated in reward, attention, and memory than minimally-processed foods or experimentally investigated the relative role of the elevated caloric density versus the flavor enhancers of ultra-processed foods in driving greater activation of these brain regions. Preliminary data showed that tastes of ultra-processed high-calorie chocolate milkshake produced greater activation in regions implicated in reward valuation (caudate, nucleus accumbens), attention/salience (precuneus), and memory retrieval (medial temporal gyrus, dorsomedial prefrontal cortex) than tastes of ultra-processed low-calorie chocolate milkshake. Aim 1 is to test the hypothesis that tastes, anticipated tastes, and images of ultra-processed foods activate reward, attention, and memory regions more than tastes, anticipated tastes, and images of minimally-processed foods, and evaluate the relative role of the higher caloric content versus flavor enhancers in engaging these regions using a 2 x 2 experimental design. Aim 2 is to test the hypothesis that ultra-processed versus minimally-processed foods promote stronger learning of cues that predict tastes of ultra-processed foods (incentive sensitization), which is important because elevated reward region response to food cues/images increases risk for future weight gain (Demos et al., 2012; Stice et al., 2015; Yokum et al., 2014). Aim 3 is to test the hypothesis that participants who show greater activation in reward/attention/memory regions in response to ultra-processed foods will consume more ultra-processed foods ad lib and show greater future body fat gain, and to establish neural fingerprints that predict ad lib ultra-processed food intake and body fat gain. Aim 4 is to test the hypothesis that participants who show stronger reward cue learning in response to ultra-processed foods will consume more ultra-processed foods ad lib and show greater ...