Attached to this page are the files needed to reproduce the meta-analysis of explore-exploit decisions. The analyses were performed according to the preregistration: https://aspredicted.org/7hc7c.pdf. The manuscript describes the PRISMA process used to identify the studies used in this analysis. Included were N=23 studies that include 13 n-armed bandit studies, 7 foraging studies, and 3 other kinds of explore-exploit tasks. The foraging and "other" studies were grouped into an "other" category of explore-exploit tasks to serve as a control condition for n-armed bandit studies. All of the studies reported exploration and exploitation decisions, or the contrast between them. These decisions were extracted and coded as exploration and exploitation for the purposes of the meta-analysis. The "Input Folder" located on OSF has the included studies and their coordinates. Studies are coded as exploration or exploitation respectively. To reproduce these results, use SDM version 6.22. Follow the instructions described in: https://www.sdmproject.com/manual/. The steps below outline how the following models were generated: 1. Figure 3: Explore and Exploit: CBMA of all explore and exploit decisions 2. Figure 4: Explore>Exploit: CBMA of the contrast between explore and exploit decisions 3. Figure 5: Explore Only, Exploit Only, Conjunction analysis 4. Figure 6: N-armed>other | Explore, N-armed>other | Exploit: CBMA of the contrast between n-armed and other tasks while making explore and exploit decisions respectively. Each model was generated in sequence using SDM 6.22. However, the steps are described in parallel below in the interest of space. 1. Generate a folder in the main directory of SDM. Use a single folder for all meta-analyses or the multimodal function will not work. 2. Copy the contents of "input folder" into this folder. 3. Select Change Meta-analysis and select your folder. 4. Copy the contents of sdm_table.txt into the SDM table editor. The "study" calls the proper, "t_thr" is the t threshold identified from the number or studies and thresholding used, "explore" and "exploit" classifies the decision, "n-armed" and "other" as the task type. "Explore>exploit" and "n-armed_other" the contrasts. 5. Follow the steps identified in the methods section and the manual to process the results. I will describe the processes specific to this project. 6. Meta analyses were run as follows: Explore and Exploit, Explore, Exploit, Explore>Exploit, N-armed>Other | Exploit, N-armed>Other | Explore. 7. The defaults were used for preprocessing parameters. 8. For Explore and Exploit, Explore, and Exploit, use the Mean feature. For Explore and Exploit, select Filter and "All". For Explore, "Explore". For Exploit, "Exploit". 9. For Explore> Exploit, N-armed>Other | Exploit, and N-armed>Other | Explore, use the Linear Model feature. Select Hypothesies as 1. For the Explore>Exploit model select "explore_exploit". For the N-armed>Other model, select "n-armed>other". To analyze the explore and exploit conditions specifically within n-armed>other, select the filter "explore" and "exploit" respectively. 10. Perform FWE correction. Default parameters were used. 11. Threshold using TFCE 12. Use Extract tool for all significant clusters. 13. Use Bias Test for each respective mask to generate funnel plots. 14. To generate conjunction image, use Multimodal function. Select the folder associated with your directory. Select the explore and exploit models you generated. 15. All thresholded images were visualized in MRIcrogl.