Abstract:  Based on previous investigations, positive and negative sign is an important factor of SNARC (Spatial-Numerical Association of Response Codes) effect. Magnitude judgments of different signed numbers are solved by an obvious sign shortcut mechanism. When both numbers are negative, there are arguments between ontogenetic hypothesis and phylogenetic hypothesis. The ontogenetic hypothesis supposes that negative numbers are mapped onto the mental number line according to their numerical value, but phylogenetic hypothesis inserts that the representation of negative numbers depends on their absolute numerical value. Whether the SNARC effect is processed in the stage of response selection or stimulus presentation is also under discussion. Although frontal lobe and parietal lobe are generally considered as the key brain regions of the SNARC effect, hemispheric dominance of this effect still needs exploration. Therefore, our research focused on four points: 1) how sign shortcut mechanism affected the SNARC effect, 2) how negative numbers represented on the mental number line, 3) how the signs of numbers affected the processing stage of the SNARC effect, 4) the key brain regions and hemispheric dominance of the SNARC effect. In the current experiment, we used modified “magnitude judgments” paradigm with ERPs recorded, to investigate how positive and negative signs influence the SNARC effect. Participants were informed the base number and the response mode before the task, then they were instructed to compare the sizes of the target numbers (–9~+9, excluded –5, 0 and +5) and the base number (–5 or +5). Two manners of keying were adopted, including congruent keying and incongruent keying. Congruent keying required participants to make “smaller” responses with the key “F” (left) and make “larger” responses with the key “J” (right). Conversely, incongruent keying required participants to make “smaller” response with the key “J” (right) and make “larger” response with the key “F” (left). Accordingly, the SNARC effect refers to the situation where the reaction times of congruent keying were shorter than those of incongruent keying. Behaviorally, different sign comparisons had higher accuracy rates than same sign comparisons. Besides, when the base number was +5, accuracy rates were higher than the condition where the base number was –5. For reaction times, responses to different sign comparisons were faster than responses to the same sign comparisons. Compared to –5, reaction time was shorter when the base number was +5. Congruent keying reacted faster than incongruent keying. In the same sign comparisons, if the base number was +5, congruent keying was faster than incongruent keying. But if the base number was –5, there was no difference between two keying types. ERP results showed that congruent keying elicited more positive P3 in response selection stage, no matter in different or same sign comparisons, which represented the SNARC effect. When base and target numbers were different signed numbers, the sign shortcut mechanism affected SNARC differently. Specificlly, when the target numbers were negative, congruent keying produced smaller N300 than incongruent keying in the stimulus presentation stage. However, when the target numbers were positive, congruent keying produced more positive LPP in the response execution stage. Traceability analysis showed that SNARC effect activated frontal lobe and parietal lobe. The negative numbers were processed with the activation of left frontal regions but positive numbers were processed with the activation of right frontal regions. Our findings suggest that: the spatial representation of negative numbers supports ontogenetic hypothesis. Positive signs and negative signs can modulate the processing stage of the SNARC effect. The spatial representation of positive and negative numbers depends on different dominant hemisphere.

Key words:  SNARC effect, sign shortcut mechanism, negative number representation, mental number line, ERP