Split Effect In Complex Addition Verification Task: Evidence From Both Behavioral And ERP Studies

Number processing and mathematical calculation is a field of research with a growing interest in cognitive neuroscience. Several phenomena have been described, but one of the interesting phenomena concerning the processing of numerical information is the effect of the manipulation of the numerical distance between the proposed and the correct solutions in a problem verification task. It is usually known as the split effect. This study explores the split effect in complex addition verification tasks using both behavioral experiments and event-related brain potentials (ERPs).In behavioral study, we used the psychological refractory period (PRP) paradigm involving a complex addition verification task and a tone-discrimination task to address the question at which stage this effect arises. In Experiment 1 the problem verification task is performed second, and we found equal split effects for the short and the long SOA. According to the central bottleneck model, this indicates that the effect arises during the response-selection or execution stage. In Experiment 2 the problem verification task was performed first. The pattern of results indicates that the split effect originates during the perceptual encoding or response-selection stage. Together, the results in our behavioral studies suggest that the split effect originates while the response is selected.In the ERP experiment, we only used the complex addition verification task. Participants were asked to verify whether the sum of the two operands was smaller than 100 or not. Behavioral data revealed that a split effect on both reaction times and error rates had showed up. Potentials related to small-split and large-split problems were analyzed to show that the perception stage is the same and that the central stage is different. Stimulus-locked lateralized readiness potentials (S-LRP) and response-locked lateralized readiness potentials (LRP-R) were computed respectively. The latency of S-LRP abruptly increased for small-split problems, thus mirroring the reaction time-split relationship. LRP-R revealed no dependency on split. These findings suggest that the processes responsible for the split effect influence processing stages that are completed before the onset of LRP. Together, the results of the ERP study indicate that the split effect arise in the response-selection stage. This is in line with the conclusion of the behavioral study.The results of the ERP study also suggest that participants may use two different strategies to verify complex inequalities. For small-split problems participants may use an exact-calculation strategy and for large-split problems, an approximate-calculation strategy. The choice between the two strategies occurred within 300ms post-stimulus presentation, and the execution process of the two strategies was different. Nevertheless, the response process was the same. Implications for understanding the brain mechanisms underlying the split effect are discussed.