In sum, the studies utilizing ERP methodology to examine the neural correlates of maltreatment have provided some evidence indicating that the experience of maltreatment, whether in the context of institutional rearing or neglect or abuse suffered at the hands of a parent, is manifested at the level of neurofunctional processes underlying memory. Although there are no studies to date specifically examining the ERP correlates of memory in children who have experienced maltreatment, the results of the ERP studies with maltreated children thus far have begun to suggest, at the very least, directions for future ERP studies of memory and maltreatment. Clearly, the experience of maltreatment early in development alters information processing systems related to recognition of emotion and attentional resources allocated to emotion recognition.
Although the studies conducted to date reveal a number of consistencies, these must be interpreted with caution given the different age ranges examined across investigations. For example, an overall pattern of hyper-responsivity in response to displays of angry facial affect among children who have experienced maltreatment has been observed. This heightened amplitude of ERP waveforms associated with angry affect is evident in the P260 waveform found in the Cicchetti and Curtis (2005) study, and in the P3b waveform observed by Pollak et al. (1997; Pollak et al., 2001). Although these two waveforms are not likely equivalent, and the ages of the participants in the studies by Pollak and colleagues (i.e., 6 to 12 years) and Cicchetti and Curtis (i.e., 28 to 36 months) are quite different, the processes underlying the heightened responsivity to angry facial affect in maltreated compared to nonmaltreated children may be parallel.
The ERP studies reviewed herein have, for the most part, tested hypotheses firmly grounded in theoretical and conceptual models of the development and functioning of emotional and attentional processes at both the behavioral and neural levels of analysis. In that sense, they represent an appropriate application of ERP methodology to answer questions tightly bound to well-explicated theoretical constructs. In addition, many of these theoretical constructions of neural processes have a great degree of face validity and "intuitive" appeal, and are grounded in comprehensive knowledge of neuroanatomical structure and functioning.
Nonetheless, it is important to note that knowledge concerning the linkage between the characteristics of ERP waveforms (i.e., morphology, amplitude) and neurobehavioral processes is far from complete. There is extensive empirical work demonstrating the neurobehavioral correlates of the most commonly observed ERP waveforms reviewed earlier in this chapter. However, the functional correlates of many waveforms observed in infants and children are less well understood (for example, the early positive component observed in Cicchetti & Curtis  and Parker & Nelson [2005a, 2005b]). Of course, part of the functionality can be inferred from the latency of a particular component. For example, a component such as the P260, given that its peak amplitude occurs, on average, 260 ms after stimulus onset, is fairly early in the neurofunctional processing stream, and thus most likely reflects some perceptual process rather than later cognitive processing.
It is also important to keep in mind that inferences made from ERP data are correlational. Accordingly, this precludes making statements about whether any given observed ERP activity is necessary for the occurrence of a particular associated emotional or cognitive process (Otten & Rugg, 2001). In addition, it is crucial to temper any interpretation of the functional significance of ERP waveforms with the reality that the association of particular patterns of ERP activity with a given experimental condition is a necessary, but not sufficient, condition for concluding that distinct neurofunctional processes are engaged. Direct behavioral assessment of the particular function of interest is necessary, as exemplified by Richards's (2003) utilization of heart-rate changes in infants to index attention.
In the context of any research involving ERPs, careful consideration of what the meaning of amplitude differences in ERP waveforms may represent is of critical importance. From a purely electrophysiological standpoint, the amplitude of a waveform is an index of the degree to which its underlying neural generators are active (represented by voltage). This, in turn, is believed to reflect the degree of engagement of the associated perceptual, cognitive, or emotional processes, with greater amplitude representing "more" activity. However, inferences concerning the "psychological" meaning of between-group (or within-group, between-condition) differences in amplitude, such as those differences found in ERP waveforms elicited by facial expressions of emotion between maltreated and nonmaltreated children, must be done with some circumspection regarding the actual "psychophysiological" meaning of amplitude.
Although between-group or between-condition comparisons of latency of ERP waveforms can provide a straightforward index of the temporal characteristics of neural processing (one of the primary strengths of ERP methodology), few if any meaningful latency differences have been found in ERP studies of the effect of maltreatment. This lack of latency effect can most likely be attributed to the typically greater interindividual variability generally found in the latencies of ERP components, which in turn reduces the ability of conventional, analysis of variance (ANOVA)-based statistical analyses to reveal such effects (i.e., due to large standard deviations).
There are also methodological limitations that are important to note in some of the ERP studies of maltreatment. Although the findings reported in the Pollak et al. (1997; Pollak et al., 2001) studies are striking, it is necessary to acknowledge that the data for these studies were collected utilizing electrodes at only three midline scalp sites (i.e., Fz, Cz, and Pz). Though current standards call for high-density arrays of 128 and even 256 electrodes, at the time these data were being collected (i.e., the mid-1990s), ERP studies utilizing only a few electrodes were state of the art. Although this does not in any way diminish the validity of the findings from these investigations, it is nonetheless important to interpret the results from these early studies with this caveat in mind.
Was this article helpful?