**4. Discussion**

The present study investigated the functional association between MA and the target-related P3 latency in the standard CPT as well as in two further conditions with variations of task demands. While in the standard condition (CPT1), a target ('X') had to be identified within a sequence of distractor stimuli, the process of selective identification was impeded in the attention-enhanced condition (CPT2) and omitted in the control condition (CPT0). This variation in task demands led to changes in P3 latency: The P3 latency was shorter in the CPT0 condition and longer in the CPT2 condition compared to the standard CPT condition. Concurrently, the correlational relationship between MA and P3 latency observed in the standard CPT condition ceased in the CPT0 condition but was even stronger in the CPT2 than in the CPT1 condition. As indicated by regression analyses, the significant amount of variance in MA explained by P3 latency in the standard CPT condition was also part of the variance explained by P3 latency in the attention-enhanced CPT2 condition. Even more importantly, the P3 latency in the attention-enhanced CPT2 condition explained an additional amount of variance of MA beyond and above the amount explained by the P3 latency in the standard CPT condition. This pattern of results strongly suggests that the functional correlation between MA and P3 latency depended on selective-attention demands.

In the standard CPT condition, attention needed to be directed to each stimulus to decide whether a target or a distractor was presented. In the CPT0 condition, only simple reactions were required so

that the demands on selective attention were minimal. Finally, in the CPT2 condition, more selective attention than in the CPT1 condition was required to identify the target not only by its characteristic of being an 'X' but also by the additional feature of being italicized. Increasing RT as well as increasing P3 latencies and amplitudes from the CPT0 to the CPT2 condition indicated increasing selective-attention demands across CPT conditions.

In the standard CPT condition, the expected negative relationship between P3 latency and MA was observed, indicating that individuals with higher MA needed less time to correctly classify, and thus identify, the target stimulus than individuals with lower MA. This result was consistent with previous studies on the relation between MA and P3 latency using the oddball paradigm [25,29–33]. As outlined above, the main difference between the CPT, as used in the present study, and the oddball task was the composition of distractor stimuli. While only one distractor stimulus was used (and frequently presented) in the oddball task, different distractor stimuli were presented in the present standard CPT. In both tasks, however, the target stimulus was infrequently presented to be identified within a series of sequentially presented distractors. The negative relation between MA and P3 latency in the standard CPT condition supports the idea that the speed of identifying targets among distractors accounts for the relation between P3 latency and MA—as it was found for the many applications of the oddball task.

Consistent with the rationale of the present study, the absence of distractors made the identification of the target vs. nontarget stimulus unnecessary. Without this demand on selective attention, however, the latency of the P3 component was no longer related to MA. The absence of an association between MA and P3 latency in the CPT0 condition (i.e., a condition without selective-attention demands) was not unexpected given the results by McGarry-Roberts et al. [22] or Troche et al. [27]. In both studies, similar results were reported for the association between MA and P3 latencies in simple reaction time tasks as in the CPT0 condition of the present study.

P3 amplitude and latency as well as RT and error rates increased from the standard CPT1 to the attention-enhanced CPT2 condition indicating that the CPT2 condition required more selective attention to identify the italic ' *X*' among distractors (including the regular 'X'). In line with our expectations, the functional relationship between MA and P3 latency became stronger in the CPT2 compared to the CPT1 condition. As indicated by regression analyses, P3 latency in the CPT2 condition not only explained an amount of variance in MA overlapping with P3 latency in the CPT1 condition but also a significant additional portion of variance in MA. This outcome provided further evidence for the notion that the selective-attention demands on target identification are decisive for the relation between MA and P3 latency.

Both RT and P3 latency proved to be similarly sensitive to our experimental manipulation. Nevertheless, the correlation between RT and P3 latency as two distinct speed measures was only weak and not significant in the CPT2 condition. These results provided additional converging evidence for the notion that RT and P3 latency represent functionally different processes [22]. Despite this apparent functional independence of P3 latency and RT, it was the most surprising result of the present study that RT was not (negatively) related to MA in any of the three task conditions. Many previous studies using the CPT analyzed errors of omission and commission regarding their relation to MA or as possible indicators of impaired attentional processes in neurologically ill individuals [38–40]. In the present study with healthy participants, both kinds of errors were extremely rare and no correlations between intelligence and errors of omission or commission could be obtained. Since errors did not differ between individuals with higher and those with lower MA and the use of different cognitive strategies was unlikely, the standard CPT met the preconditions for a systematic relationship between RT and MA [1,2]. It should be noted, however, that a lacking negative relation between MA and RT is not unusual in the field of mental chronometry [41,42]—even though the majority of studies revealed a weak to modest relationship between MA and RT [3].
