<p>First-graders’ patterning skills are a robust predictor of later mathematics achievement, underscoring the need to support these skills early, particularly for students at risk of developing mathematical difficulties. Although previous studies have demonstrated the general benefits of patterning instruction, they have rarely examined qualitative shifts in the processes involved. We investigated whether a one-to-one support program could (a) shift pattern-recognition processes (as inferred through eye movements), (b) lower error rates, and (c) shorten response times in first-grade students at risk of developing mathematical difficulties (RMD), relative to at-risk and typically achieving peers who received no extra instruction. In this study, 34 first-graders completed a standardized mathematics test and were assigned to an RMD support group, an untreated RMD control group, or an untreated nonRMD control group. All students completed 21 repeating-pattern tasks before and after the 6-week intervention, with their eye movements being recorded. The RMD support group showed the most significant redistribution in their pattern-recognition processes, shifting from focusing on one unit of repeat or an element-by-element approach to more relational thinking about patterns. Control-group changes were less instructionally desirable: the untreated RMD control group shifted toward an inefficient element-by-element approach that reduced errors but increased response times. Error rates decreased in both RMD groups, suggesting no treatment effect on performance. Because these process shifts were invisible in error rates alone, the study highlights the value of combining eye-movement data with traditional performance measures when evaluating early mathematics interventions.</p>