3.2.2. Frontal Activity

The upper panel of Figure 4 shows lure-evoked ERPs at FCz. A large negative peak seems to be evoked by the first color lure at about 300 ms. However, the scalp map of these data suggests that this peak may stem from volume-conducted posterior negativity. In contrast, negative peaks with fronto-central focus are visible with 2nd and 3rd lures, though at later latencies. In order to enhance the weights of location-specific activity and ge<sup>t</sup> rid of volume conducted posterior negativity, current source densities (CSDs) [32] were computed (lower panel of Figure 4). Indeed, as the maps in Figure 4 show, current-sinks had a focus at or near FCz, with their amplitudes appearing to vary between serial positions: negligible activity with the first lure, distinct but relatively late activity with the second lure (400–500 ms after lure onset), somewhat earlier activity (300–400 ms) with the third lure.

**Figure 4.** ERPs and current source densities evoked by the lures. Data are grand means across participants, recorded from the fronto-central midline site FCz. The upper panel displays the ERPs, and the lower panel displays current source densities, i.e., ERP data from which ERP data from surrounding sites were subtracted. Depicted are differences between lure-trials and corresponding epochs of no-lure trials. Unit on x-axis is milliseconds, time-point zero is lure onset. Unit on y-axis is microvolts in the upper panel and microvolts per square meter in the lower panel, negative values are plotted upwards. Waveforms evoked by the 1st lure are shown as solid lines, by the 2nd lure as dashed lines, and by the 3rd lure as dotted lines. Data evoked by color lures and digit lures are plotted with blue and grey lines, respectively. The scalp maps show the view on the head (120◦) from above. Recording sites are denoted by the small circles. Displayed are topographic distributions at the indicated latencies for each of the six conditions. In the upper panel, blue denotes negative polarity, red positive polarity, and scale is ±2 μV. In the lower panel, blue denotes negative sinks, red denotes positive sources, and scale is ±15 μV/m2.

To quantify these impressions, an ANOVA was computed on mean CSD amplitudes of the 300–400 ms and 400–500 ms epochs, with the factors Epoch (early, late), Serial Position (1st, 2nd, 3rd lure), Lure Type (digit, color), and Lure Stream (left, right). The main effect of Serial Position (*F*2,26 = 3.8, *p* = 0.047) and the interaction of Epoch × Serial Position (*F*2,26 = 3.9, *p* = 0.043) were further explored by computing pairwise ANOVAs on lure #1 vs. #2 and on #2 vs. #3 separately for the two epochs. Negativity increased from lure #1 to #2 in the late epoch (*F*1,13 = 6.6, *p* = 0.02; early epoch: *<sup>F</sup>*1,13 = 0.6, n.s.) and from lure #2 to #3 in the early epoch (*F*1,13 = 7.3, *p* = 0.02; late epoch: *<sup>F</sup>*1,13 = 0.1, n.s.). Different from what we had expected, there was hardly any difference between digit and color lures. The only effect of Lure Type was a moderation of the just-mentioned Epoch × Serial Position interaction, Epoch × Serial Position × Lure Type *<sup>F</sup>*2,26 = 4.6, *p* = 0.02, which appeared to reflect a tendency with 3rd lures where negativity tended to decrease from early to late epoch with digit lures, *<sup>F</sup>*1,13 = 4.3, *p* = 0.06, not so with color lures, *<sup>F</sup>*1,13 = 0.0, n.s. There were no reliable differences between left and right streams (all effects of Lure Stream *<sup>F</sup>*1,13 ≤ 3.6, *p* ≥ 0.08).

#### *3.3. ERP Reflections of Lure E*ff*ects on T1 Processing*

T1-evoked ERPs were computed from about 120 trials per participant (150 trials per cell minus about 20% trials with incorrect responses) minus artifact-affected trials.

T1-evoked contralateral–ipsilateral differences from posterior–lateral sites |PO7–PO8| are displayed in Figure 5. There is a large N2pc, peaking at about 220 ms. Lure effects were measured by computing the differences between lure trials and no-lure trials in the mean amplitudes at 175–275 ms for each of the four lure conditions and submitting these difference values to an ANOVA with the factors Lure Type (color, digit) and Lure-T1 Relation (same side, other side).

**Figure 5.** Contralateral–ipsilateral ERP differences evoked by T1. Data are grand means across participants, recorded from left and right posterior sites PO7 and PO8. Unit on x-axis is milliseconds, time-point zero is lure onset. Unit on y-axis is microvolts, negative voltage is plotted upwards. Waveforms evoked in no-lure trials are black, from color-lure trials blue, and from digit-lure trials grey. Trials where lures were in the same stream as T1 are denoted by solid lines (grey and blue) and trials where lures were in the opposite stream are denoted by dashed lines. The scalp map shows the view on the head (120◦) from above. Recording sites (small circles) are depicted on one hemisphere only because the contralateral–ipsilateral differences were pooled across left and right sides. Blue is contralateral negativity, red is positivity. Scale is ±7 μV. Displayed is the topographic distribution of N2pc evoked in no-lure trials at the peak latency of N2pc (222 ms).

T1-evoked N2pc reliably increased when color lures had preceded on the other side. (Main effect of Lure-T1 Relation *<sup>F</sup>*1,13 = 19.8, *p* = 0.001, modified by the interaction of Lure-T1 Relation × Lure Type *<sup>F</sup>*1,13 = 5.9, *p* = 0.03, resolved to a simple effect of Lure-T1 Relation with color lures, *<sup>F</sup>*1,13 = 55.8, *p* < 0.001, and absence of such effect with digit lures, *<sup>F</sup>*1,13 = 0.3, n.s.). The main effect of Lure Type was not significant (*F*1,13 = 1.0, n.s.).

Confirming the ANOVA results, *t* tests of the four lure conditions against no lures yielded a significant increase of negativity for color lures in the other stream, *t* = −4.6, *p* < 0.001, and no significant differences for any of the three other conditions, *t* ≤ |1.2|, *p* ≥ 0.24.

#### *3.4. ERP Reflections of Lure E*ff*ects on T2 Processing*

T2-evoked ERPs were computed from 28 trials per participant (50 trials per cell minus about 33% trials with incorrect responses minus artifact-affected trials).

T2-evoked contralateral–ipsilateral differences from posterior–lateral sites |PO7–PO8| are displayed in Figure 6, separately for T2 on the same side as T1 and T2 on the other side from T1. Since T2 followed on T1 with a lag of 3, = 330 ms, the leftmost time-point of Figure 6 (−100 ms) is time-point 230 ms of Figure 5 (except that Figure 5 also includes trials where T2 followed T1 at lag 1 whereas Figure 6 includes trials with lag 3 only). At that time, T1-evoked N2pc just reaches its peak in Figure 5. Thus, Figure 6 starts with the decreasing slope of T1-evoked N2pc. This decrease is plotted with same polarity (descending towards positivity) when T2 and T1 are on the same side (left panel of Figure 6) and with opposite polarity in the other case (ascending towards negativity; right panel).

**Figure 6.** Contralateral–ipsilateral ERP differences evoked by T2. Data are grand means across participants, recorded from left and right posterior sites PO7 and PO8. Unit on x-axis is milliseconds, time-point zero is lure onset. Unit on y-axis is microvolts, negative voltage is plotted upwards. Waveforms evoked in no-lure trials are black, from color-lure trials blue, and from digit-lure trials grey. Trials where lures were in the same stream as T2 are denoted by solid lines (grey and blue) and trials where lures were in the opposite stream are denoted by dashed lines. The left panel displays data where T1 was in the same stream as T2, and the right panel displays data where T1 was in the other stream. Time-point −100 ms is approximately the peak of that preceding T1. The scalp maps show the view on the head (120◦) from above. Recording sites (small circles) are depicted on one hemisphere only because the contralateral–ipsilateral differences were pooled across left and right sides. Blue is contralateral negativity, red is positivity. Scale is ±8 μV. Displayed is the topographic distribution of N2pc evoked in no-lure trials at the peak latencies of N2pc (282 ms in same-stream-as-T1 trials, 260 ms in opposite-stream trials).

There is a distinct N2pc when T2 was preceded by T1 in the other stream (right panel of Figure 6), peaking at about 250 ms, and a similar though apparently smaller component is visible when T2 was preceded by T1 in the same stream (left panel of Figure 6). To quantify these impressions, N2pc was measured by computing mean amplitudes 200–300 ms.

First, T2-evoked N2pcs were compared between same-side T1 and other-side T1 in an ANOVA on no-lure trials with the one factor T1-T2 Relation (same side, other side). Indeed, N2pc was larger after other-side T1 than after same-side T1, *<sup>F</sup>*1,13 = 13.8, *p* = 0.003. Then, lure effects were measured by computing the differences between lure trials and no-lure trials, separately for each of the four lure conditions and for T1-T2 same-side and other-side trials, and submitting these difference values to an ANOVA with the factors Lure Type (color, digit), Lure-T2 Relation (same side, other side) and T1-T2 Relation (same side, other side). In this ANOVA on difference values, the constant term was different from zero, *<sup>F</sup>*1,13 = 7.2, *p* = 0.02, indicating that there was an overall N2pc difference between lure and no-lure trials which, as Figure 6 shows, was an overall reduction of N2pc by the presence of lures. None of the factors had significant influence on this lure effect, though, all *<sup>F</sup>*1,13 ≤ 2.8, *p* ≥ 0.12. When nevertheless testing each of the eight lure conditions separately against the no-lure condition, reductions were significant when digit lures preceded on the same side and also T1 was on that side (*t* = 2.3, *p* = 0.04; solid grey line in left panel of Figure 6) and when color lures preceded on the same side and T1 was on the other side (*t* = 2.7, *p* = 0.02; solid blue line in right panel of Figure 6).
