600 mm

FIGURE 18.2 Grand mean ERPs over frontal and fronto-central electrodes recorded during tempo encoding in the six-interval condition (upper part of the figure) and in the three-interval condition (lower part of the figure). (Adapted from Pfeuty, M., Ragot, R., and Pouthas, V., Psychophysiology, 40, 1-8, 2003.)

congruent with other reports (Dirnberger et al., 2000; Ferrandez and Pouthas, 2001). In our study, CNV amplitude in a duration discrimination task was lower in middle-aged adults (50 years) than in the young ones (20 years) when measured at frontal sites. However, it was higher when measured at more central sites, as shown in Figure 18.3. The finding of diminished frontal activity in older subjects suggests a process of selective cortical aging and possible cellular loss, which may be linked to performance deficits. We proposed that modifications of the topography of electrical activity in middle-aged and elderly adults could reflect a reorganization

FIGURE 18.3 ERP waveforms at F1, F2, FCz, C1, and C2 sites for young adults (thin line) and middle-aged adults (thick line) in a duration discrimination task. (Adapted from Ferrandez, A.M. and Pouthas, V., Neurobiol. Aging, 22, 645-657, 2001.)

process, with old adults using more posterior areas to compensate for structural deficits in more frontal sites (Raz et al., 1997).

The amplitude of the CNV has also been shown to be significantly reduced in Parkinson's disease (PD) patients compared with age-matched controls (Ikeda et al., 1997; Praamstra et al., 1996; Pulvermuller et al., 1996; Wascher et al., 1997). Temporal performance of these PD patients is known to be impaired (Malapani et al., 1998; Pastor et al., 1992). Ikeda et al. (1997) found that the level of CNV reduction in PD patients, particularly over the frontal areas, was directly related to the severity of disease. Concurrently with temporal performance improvement, the amplitude of the CNV has been shown to increase following treatment with levodopa (Amabile et al., 1986; Oishi et al., 1995) or when patients receive bilateral subthalamic nucleus stimulation, as shown in a study by Gerschlager et al. (1999). These authors examined changes in the CNV, recorded from patients with Parkinson's disease when on and off bilateral subthalamic nucleus stimulation, and compared these data with the CNV of healthy control subjects. Without subthalamic nucleus stimulation, Parkinson's disease patients showed reduced CNV amplitudes over the frontal and fronto-central regions compared with control subjects, whereas with bilateral subthalamic nucleus stimulation, CNV amplitudes over the frontal and fronto-central regions were significantly increased. These results suggest that subthalamic nucleus stimulation in PD improves the cortical activity that underlies processing of temporal information (for additional details of the effects of subthalamic stimulation on interval timing in PD patients, see Malapani and Rakitin, this volume).

18.2.3 CNV Resolution and Temporal Processing

The amplitude of the CNV is not the only important parameter for assessing relationships between cortical activity and temporal judgments. Indeed, the resolution of this slow wave, i.e., its return to baseline, has also been taken into account. In experimental paradigms specifically designed to study time estimation, it has been shown that the return of the negativity to baseline started largely before the imperative stimulus (S2) and the motor response that are most often temporally close together (Ladanyi and Dubrovsky, 1985; Macar and Vitton, 1982; Ruchkin et al., 1977). Macar and Vitton (1982) asked whether the early resolution (ER) has to be referred to processes specifically involved in time estimation paradigms. Their experiment was aimed at checking whether the CNV presented an early resolution in a time discrimination paradigm, which dissociated cognitive timing processes from motor mechanisms more clearly than the S1-S2 paradigm more commonly used. In the latter, the authors observed a classical CNV during the delay that is interpreted as an index of expectancy. By contrast, in the task in which the subject had to collect temporal information, the CNV was followed by an early and prolonged resolution of negativity in the second half of the interval. Macar and Vitton (1982) concluded that this type of event-related potential (ERP) component, identified as a CNV-ER, appeared as a correlate of cognitive processes pertaining to the collection of temporal information.

An important issue is to determine the moment in time of the shift from negative to positive components of the CNV. The next three studies that we will describe address this question. Ladanyi and Dubrovsky (1985) investigated the CNV waves of subjects classified according to their performance in time estimation tasks in accurate and nonaccurate estimators. As mentioned above, accurate estimators have CNVs of lower amplitude and show a slower rise time to peak negativity than subjects with poor time estimation abilities. Most importantly for our purpose, the results revealed that subjects with a high degree of accuracy in time estimation tasks exhibited a faster resolution of the negativity. Moreover, the data also showed a positive correlation between delay in resolution of the CNVs and the degree of error in the subjects that overestimate time periods; i.e., the more the subjects overestimate time lapses in the behavioral tests, the more prolonged were their CNVs waves. The authors assumed that the development and resolution of the CNV may, in part, reflect time estimation processes in the nervous system. They concluded that CNV resolution is most likely related to the conscious decision to respond after the formation of the temporal judgment.

Another example can be found in a study by Ruchkin et al. (1977). Subjects had to reproduce a target interval of 900 msec after a click. The authors examined average waveforms synchronized to the click for three categories of interval reproduction times, i.e., 800 msec (752 to 828), 900 msec (892 to 908), and 1000 msec (972 to 1048). Average waveforms synchronized to the click for 800, 900, and 1000 msec evidenced a covariation between subjects' temporal judgments and the latency of the negative-to-positive amplitude shift of the concomitantly recorded ERPs. These results suggest that the differences between the latencies reflect differences in timing of the cognitive processes associated with the formation of a temporal judgment.

These data sets suggest that CNV resolution may relate, at least in part, to the conscious decision to respond after the formation of a temporal judgment. Therefore, the resolution of the CNV is not necessarily time-locked to an imperative stimulus. This is particularly well exemplified by results of one of our recent studies (Pouthas et al., 2000). We recorded ERPs in a matching-to-sample task in which subjects had to decide whether the duration of an light-emitting diode LED illumination was the same as or different from a standard duration (700 msec) previously memorized. As demonstrated in Figure 18.4, the CNV resolution occurred after the LED switched off. However, the moment of resolution occurrence depended on whether the duration of the standard had elapsed. For test durations shorter than the standard (i.e., 490 and 595 msec), the resolution intervened later after LED switch off, i.e., around 200 msec, compared to what happened with durations longer than the target (805 and 910 msec), i.e., around 100 msec, as shown in Figure 18.4. We proposed the following explanatory hypotheses to account for these results. On the one hand, slower resolution of the CNV would reflect the fact that subjects were waiting for the standard duration to elapse before making a decision. On the other hand, faster resolution would reflect the fact that after the standard duration was over, subjects had formed their judgment and anticipated the LED switch off.

This section has revealed that changes in the accuracy and precision of temporal performance are accompanied by reliable changes in brain activity as measured by event-related brain potentials, particularly by the CNV. As stressed by Macar et al. (1999), these changes should occur over the cerebral regions specifically concerned

FIGURE 18.4 ERP waveforms recorded on the FCz electrode for the five stimulus durations (upper part of the figure). Diagram showing the delay between the moment in time of the LED switch off and the moment in time of the CNV return to baseline, i.e., zero-crossing (lower part of the figure). (Adapted from Pouthas, V., Garnero, L., Ferrandez, A.M., and Renault, B., Hum. Brain Mapping, 10, 49-60, 2000.)

FIGURE 18.4 ERP waveforms recorded on the FCz electrode for the five stimulus durations (upper part of the figure). Diagram showing the delay between the moment in time of the LED switch off and the moment in time of the CNV return to baseline, i.e., zero-crossing (lower part of the figure). (Adapted from Pouthas, V., Garnero, L., Ferrandez, A.M., and Renault, B., Hum. Brain Mapping, 10, 49-60, 2000.)

with temporal processing. ERPs reflect the rapidly changing electrical activity in the brain evoked by a stimulus or a cognitive event and permit us to discriminate between different stages in information processing. It is difficult, however, to determine the neural sources of ERP components. Therefore, in the following section, we report data from neuropsychological studies and from a combination of PET and ERP analyses that provide some evidence on the major contributing sources of the CNV.

0 0

Post a comment