Maybe, but looking at the graph the peak latency looks to be about 360 milliseconds, whereas the measure given in the Neuronetrix study (on which the healthy control from the graph is based) records P300/P3b latency as 396 milliseconds. On the graph, 396 milliseconds seems to correspond to the end of the peak wave rather than its absolute peak, right where it forms a sort of "shoulder" before a sharp drop-off.
Also, if you look at the baseline on the graph, it's difficult to identify a single peak, which may account for why the latency measure appears to correspond to the "shoulder" rather than the absolute peak. The "shoulder" on the 2-73 line does appear closer to the healthy control line than to the baseline.
In any case, the Neuronetrix study shows greater statistical significance for P300/P3b amplitude than latency in discerning between healthy control and mild AD (reversing the tendency in older studies to regard latency as the more reliable measure), which could account for why Anavex has emphasized P300/P3b amplitude over latency. The measures Anavex has emphasized for latency (the target detection tasks on slide 24) also achieved greater statistical significance in the Neuronetrix study than did P300/P3b latency.