Werner Sommer

Human brain potential correlates of face encoding into memory

ERPs elicited by photographs of unfamiliar faces were shown to be predictive for their later recognition. We established that these ERP differences were unrelated to fluctuations of attention or other non-specific factors during perceptual processing. Therefore they may be interpreted as manifestations of brain processes that correlate with memory encoding. The scalp topography of this ERP difference was bipolar with greater electrical positivity at frontal and greater negativity at parieto-temporal scalp sites. This topography appears to contrast with the more uniformly positive differences reported for verbal stimuli but is in accord with what might be expected for faces and complex visual stimuli.

Metamemory, distinctiveness, and event-related potentials in recognition memory for faces

A neglected topic in metamemory research is the ability of subjects to predict their own recognition performance for faces. We investigated whether subjects can make such judgments of learning (JOL) for unfamiliar faces and whether JOLs relate to facial distinctiveness, a powerful determinant of face recognition. One group of subjects made JOLs, and a second group rated the same faces for distinctiveness; subsequently, both groups tried to recognize these faces among new faces. There was significant prospective metamemory for faces that appeared to be based on facial distinctiveness. Both prospective metamemory and distinctiveness ratings related to long-lasting effects in event-related brain potentials (ERPs), closely resembling an ERP component that predicted face recognition. Therefore, the brain processes underlying JOLs, distinctiveness, and recognition memory for faces appear to be intimately related.

Covert effects of alcohol revealed by event-related potentials

Contradictory evidence as to the effects of alcohol on early information processing stages has been obtained from behavioral and psychophysiological investigations. In the present study, choice reaction times, error rates, and event-related potentials (EEPs) were recorded in a task in which variations in stimulus discriminability and of the (task irrelevant) correspondence between stimulus location and response location were orthogonally combined. Both discriminability and stimulus-response correspondence affected reaction time and electrophysiological Chronometrie measures as expected. However, no behavioral effects of alcohol were observed, possibly because of strategic adjustments. Psychophysiological Chronometrie measures indicated that alcohol leaves the initial flow of perceptual evidence to motor stages unimpaired, whereas it appears to increase the duration of stimulus evaluation. Interestingly, a number of alcohol effects appeared in the EEP amplitudes. Decrements in early EEP components indicate alcohol-induced impairments of involuntary visual attention and/or the automatic stimulus location-dependent activation of response channels. In contrast, a strong enhancement of a late slow-wave component under alcohol may reflect the investment of processing resources in order to maintain normal performance levels.

Differential effects of voluntary expectancies on reaction times and event-related potentials: Evidence for automatic and controlled expectancies

Expectancy has been used to explain the effects of stimulus sequences both on reaction times (RTs) and on the P300 component of the human event-related potential. However, there are conflicting views about the control obtainable over these underlying expectancies. We compared the effects of voluntary expectancies for stimulus changes or repetitions in random tone series on RTs and the P300. Ss responded according to either stimulus identity (Experiment 1) or stimulus sequence (Experiment 2). In both experiments RTs were strongly affected by event expectedness. P300 amplitude, on the other hand, was affected (as a trend) only in Experiment 2. The results suggest that there are at least 2 types of expectancy, one that is largely automatic and inflexible, reflected in P300 amplitude, and a second, controlled process that is reflected mainly in RT. The latter type of expectancy appears to affect processing stages beyond stimulus evaluation and classification.

Contributions of stimulus encoding and memory search to right hemisphere superiority in face recognition: Behavioural and electrophysiological evidence

The relative contributions of stimulus encoding and memory search to right hemisphere (RH) superiority in face recognition were investigated in a memory search task with lateralized probe stimulus presentation. RH superiority in reaction time and latency of a late positive ERP component was independent of stimulus degradations, but increased as a function of memory set size. These results suggest equal efficiency of the hemispheres in stimulus encoding, but RH superiority in memory search. In addition, a long-lasting ERP negativity contralateral to the stimulated hemifield is discussed with reference to models of hemispheric processing in hemifield studies.

Covert face recognition in prosopagnosia: a dissociable function?

Covert face recognition was investigated in a patient with prosopagnosia without object agnosia. This patient performed well in various face processing tasks like expression analysis and feature processing and had relatively preserved semantic knowledge about persons, but was slightly impaired in the visual matching of unfamiliar faces. In face-name paired-associate relearning task, covert face recognition was demonstrated to be above-chance. However, as this task cannot be meaningfully applied to control subjects, results do not necessarily indicate that the degree of covert face recognition is normal. In fact, in contrast to control subjects, the patient showed significantly reduced associative priming of names by face primes as compared to name primes, suggesting a quantitative reduction of covert face recognition. It is argued that these results support the view that overt and covert face recognition are brought about by the same functional system (Farah, OReilly and Vecera, 1993).

Event-related potentials and models of performance asymmetries in face and word recognition

Performance asymmetries in divided visual field studies may be ascribed either to hemispheric differences in processing efficiency or to the costs of interhemispheric transfer towards the superior hemisphere. In order to distinguish between these alternatives, event-related potentials (ERPs) were recorded while subjects had to recognize laterally presented faces or words. As expected, behavioural left- and right-field advantages were observed for faces and words, respectively. Regardless of stimulus type, the ERPs displayed a sustained temporo-parietal negativity over the hemisphere stimulated directly via the contralateral hemifield. Both this hemifield-dependent negativity (HDN) and the performance asymmetries diminished to insignificance when the same stimuli were presented but subjects simply made a left-right decision about stimulus location. We conclude that the HDN is no obligatory, stimulus-bound ERP component but depends on lateralized cognitive processing. The stimulus-unspecific and time-invariant topography of the HDN might indicate that it relates to the allocation of processing resources to the directly stimulated hemisphere. The findings suggest that both faces and words were processed predominantly in the directly stimulated hemisphere, supporting an efficiency explanation of the performance asymmetries.

The expectancies that govern the p300 amplitude are mostly automatic and unconscious

of the original article: P3 is the most prominent of the electrical potentials of the human electroencephalogram that are sensitive to psychological variables. According to the most influential current hypothesis about its psychological significance [E. Donchin’s], the “context updating” hypothesis, P3 reflects the updating of working memory. This hypothesis cannot account for relevant portions of the available evidence and it entails some basic contradictions. A more general formulation of this hypothesis is that P3 reflects the updating of expectancies. This version implies that P3-evoking stimuli are initially unexpected but later become expected. This contradiction cannot be resolved within this formulation. The alternative “context closure” hypothesis retains the concept of “strategic information processing” emphasized by the context updating hypothesis. P3s are evoked by events that are awaited when subjects deal with repetitive, highly structured tasks; P3s arise from subjects’ combining successive stimuli into larger units. The tasks in which P3s are elicited an accordingly be classified in terms of their respective formal sequences of stimuli. P3 may be a physiological indicator of excess activation being released from perceptual control areas. Commentary on Emanuel Donchin and Michael G. H. Coles (1988). Is the P300 component a manifestation of context updating? BBS 11:357-374. Abstract of the original article: To understand the endogenous components of the event-related brain potential (ERP), we must use data about the components’ antecedent conditions to form hypotheses about the information-processing function of the underlying brain activity. These hypotheses, in turn, generate testable predictions about the consequences of the component. We review the application of this approach to the analysis of the P300 component. The amplitude of the P300 is controlled multiplicatively by the subjective probability and the task relevance of the eliciting events, whereas its latency depends on the duration of stimulus evaluation. These and other factors suggest that the P300 is a manifestation of activity occurring whenever one’s model of the environment must be revised. Tests of three predictions based on this “context updating” model are reviewed. Verleger’s critique is based on a misconstrual of the model as well as a partial and misleading reading of the relevant literature.of the original article: To understand the endogenous components of the event-related brain potential (ERP), we must use data about the components’ antecedent conditions to form hypotheses about the information-processing function of the underlying brain activity. These hypotheses, in turn, generate testable predictions about the consequences of the component. We review the application of this approach to the analysis of the P300 component. The amplitude of the P300 is controlled multiplicatively by the subjective probability and the task relevance of the eliciting events, whereas its latency depends on the duration of stimulus evaluation. These and other factors suggest that the P300 is a manifestation of activity occurring whenever one’s model of the environment must be revised. Tests of three predictions based on this “context updating” model are reviewed. Verleger’s critique is based on a misconstrual of the model as well as a partial and misleading reading of the relevant literature. The expectancies that govern the P300 amplitude are mostly automatic and unconscious W. Sommer, H. Leuthold, & J. Matt Humboldt University at Berlin, Institut für Psychologie, 10117 Berlin, Germany. werner.sommer@rz.hu-berlin.de Abstract: We argue that probability effects on P300 amplitude are the product of an automatic frequency detector not subject to voluntary control and relatively inaccessible to consciousness. “Expectancies” related to P300 therefore appear to be passive, perceptual ones. If probability-based expectancies do become conscious, they are inversely related to P300, supporting the view of Donchin & Coles (1988). We argue that probability effects on P300 amplitude are the product of an automatic frequency detector not subject to voluntary control and relatively inaccessible to consciousness. “Expectancies” related to P300 therefore appear to be passive, perceptual ones. If probability-based expectancies do become conscious, they are inversely related to P300, supporting the view of Donchin & Coles (1988). A central point in the controversy between Verleger (1988) and Donchin & Coles (1988) concerns the direction of the relationship between P300 amplitude and subjective probability. The traditional position, taken by Donchin & Coles, is that P300 is inversely proportional to the expectedness of a stimulus whereas Verleger suggests it is directly proportional to the “awaitedness” of the eliciting stimulus. If the expectancies that affect P300 were conscious, one might use subjective reports to test the alternative hypotheses about the direction of the relationship. Verleger explicitly conceives of awaitedness as a conscious process (p. 403) whereas Donchin & Coles are not clear on that point. Johnson and Donchin (1982), however, and especially Johnson (1988a), have suggested that the expectancies induced by the global stimulus probability in a given experimental situation or by local variations of the stimulus pattern in a random series (the stimulus sequence) affect the P300 amplitude via an automatic – and unconscious (Karis et al. 1983) – frequency detecting mechanism. Several experiments have tested whether subjective probability is conscious and, if so, how it relates to P300 amplitude. Most of these studies have concentrated on the effects of the various sequences in randomly ordered stimulus series. Chesney and Donchin (1979) and Munson et al. (1984) used predictions about forthcoming stimuli as measures of conscious expectancies. Whereas Chesney and Donchin did not find a relationship between predictions and the P300 amplitude, Munson et al. found stimulus repetitions to elicit larger P300s when disconfirming a prediction, that is, when they had been unexpected. For stimulus alternations, however, the prediction was not related to P300 amplitude. Sommer et al. (1990) used retrospective reports about the expectedness of already presented stimuli. In their study, short random series of high and low tones were presented to which the subject had to perform a choice reaction time (RT) response. After about eight tones, the series was interrupted and the subjects reported whether they had expected the presented tone, the alternative tone, or no particular tone. Although there were systematic and graded effects of 16 different sequences of stimulus repetitions and alternations on both P300 amplitude and RT, this did not hold for the verbally reported expectancies, which were highly uniform over most sequences. There were only two exceptions: when three consecutive stimulus alternations were followed by another alternation, it was relatively Continuing Commentary 150 BEHAVIORAL AND BRAIN SCIENCES (1998) 21:1 expected; when they were followed by a repetition, however, it was rather unexpected. The contrast between the systematic relationship of the stimulus sequence to P300 amplitude (and RTs) and the isolated effect of just two particular sequences on conscious expectancy indicates a very limited relationship between P300 amplitude and conscious expectancies. As regards the relationship between P300 amplitude and the verbally reported expectancies, the P300 was larger when the stimulus was expected, but only after first-order alternations, whereas after stimulus repetitions there was a small, nonsignificant amplitude difference in the opposite direction. An even more direct test of a possible relationship between conscious expectancies and P300 amplitude was performed in two experiments where subjects were asked to actively expect particular events (Matt et al. 1992). In the first experiment, subjects performed choice responses to random series of equiprobable tones. In separate conditions, subjects either prepared for stimulus repetitions or stimulus alternations or they did not hold particular expectancies. The expectancy manipulation had a strong effect on RTs, which were shorter for expected than unexpected events. There was no effect, however, on P300 amplitude. In a second experiment, a similar procedure was used except that the subject responded not to individual stimuli but to stimulus sequences. One key was pressed when the same stimulus as before was presented and the other key was used for stimulus alternations. RTs (and error rates) were generally increased as compared to the first experiment and there were similar effects of the expectancy instructions. This time we also observed a trend towards expectancy effects on the P300 amplitude. When no stimulus was voluntarily expected, P300 amplitude was smaller to repetitions than to alternations, as is often observed; this repetition effect was increased when repetitions were expected and abolished when alternations were expected. Most recently, Sommer et al. (submitted) have investigated sequential effects in two sensory modalities when the responsestimulus interval (RSI) was reduced from a conventional 500 msec to a mere 40 msec. Whereas sequential effects in RTs changed from an expectancy-indicating cost-benefit pattern to a cost-only pattern, the cost-benefit pattern in P300 amplitude was unaffected. Provided that decreasing RSI depletes the available time resources and given that controlled processes are time-costly, these results support the hypothesis that sequential effects in P300 amplitude are the product of an automatic process. Three conclusions can be drawn from these findings. First, consistent with Johnson’s (1988a) view, probability effects on P300 are the result of an automatic process over which very little control can be exerted. Second, most of the product

Operant conditioning of P300

This study demonstrates that operant conditioning may increase the P300 component of the event-related potential above a level obtained without contingent training. An experimental group of subjects was rewarded for producing large P300 amplitudes and was compared with a yoked control group which was rewarded on a random basis. During training the experimental subjects increased both the amplitude of the P300 and of a subsequent frontal negative slow wave relative to the control group. These training effects were independent of prestimulus potential shifts and occurred likewise for target and nontarget stimuli.

Reaction Time Improvements with Practice in Brain-Damaged Patients

This study investigated practice effects on reaction time (RT) in patients with unilateral brain damage and age-matched controls. Subjects performed a Sternberg-type stimulus recognition task. Both RT and RT variability decreased in all subjects from a short initial session to a second session, 1 to 3 days later. Within the second session, however, RT remained constant in the controls, whereas it was further reduced in the patients. These findings argue against the assumption that brain-damaged patients generally show smaller practice effects on RT than controls, and instead suggest that it may be essential to provide patients with sufficient practice opportunities. To locate the processing stage where practice is effective, analyses were performed also on the intercept and slope parameters of the RT function. It is argued that the differential practice effects were due to the speeding of perceptual and/or response-related stages, rather than gains in memory search speed.