- Email: [email protected]
Aging and respond-withhold tasks: Effects on sustained potentials, P3 responses and late activity
E/ectroencephalography and clinical Neurophvsiologv, 1984, 5 8 : 1 3 0 - 1 3 9 Elsevier Scientific Publishers Ireland. Ltd.
130
Experimental Section AGING AND RESPOND-WITHHOLD TASKS: EFFECTS ON SUSTAINED POTENTIALS, P3 RESPONSES AND LATE ACTIVITY t J.A. PODLESNY 2, R.E. D U S T M A N 3 and D.E. S H E A R E R
Neuropsychology Research Laboratory', Veterans Administration Medical Center, Salt Lake Ci(v, UT 8414~ ¢U.S.A.) (Accepted for publication: January 11, 1984)
Reported effects on sustained potential 4 (SP, or CNV) and P3 responses of young adults during active respond-withhold tasks have been interpreted as consistent with the hypothesis that inhibitory cortical activity may be reflected by greater positive late potentials (Howard et al. 1980). These tasks require subjects to produce motor responses during some trials and withhold responses during other randomly presented trials. Presumably, since subjects must inhibit motor activity during withhold trials, their ERPs may reflect neural inhibition. P3 amplitude was reportedly larger during withhold trials than during respond trials (Karlin et al. 1970; Simson et al. 1977). Elderly people are believed to have weakened inhibitory function (Jakubczak 1973; Scheibel and Scheibel 1975; Beck 1978; Frolkis and Bezrukov 1979; McGeer and McGeer 1980), possibly related to reduced neurotransmitter activity in late life
1 This research was supported by the Medical Research Service of the U.S. Veterans Administration and by the U.S. National Institutes of Health Grants AG05126 and AG00568. 2 J.A. Podlesny is currently with the Laboratory Division, Federal Bureau of Investigation, Washington, DC 20535, U.S.A. 3 Requests for reprints should be addressed to Dr. Robert E. Dustman, Neuropsychology Research Laboratory (151A), Veterans Administration Medical Center, Salt Lake City, UT 84148, U.S.A. 4 We have used the term 'sustained potential' rather than 'contingent negative variation' because studies of functional and topographic differences in the early and late portions of the wave form (Rohrbaugh et al. 1976, 1980) suggest 'that it is constituted largely of separate noncontingent elements related to sensory and motor processes' (Rohrbaugh et al. 1980, p. 1165). 'Sustained potential' is a relatively neutral term which describes the sustained quality of the observed potentials without suggesting contingency which is less obvious.
(Agrawal and Himwich 1970; Beck 1978; Adolfsson et al. 1979; McGeer and McGeer 1980; Pradhart 1980). We and others have reported that evoked potentials recorded from elderly subjects in a passive state (no task involved) often show characteristics which suggest reduced inhibitory function, i.e., enhanced wave amplitude (Dustman and Beck 1969; Straumanis et al. t973; Dustman and Snyder 1981) and a relative inability of the visual system to differentiate patterned from unpatterned flashes (Dustman et al. 1981). More recently, we studied aging effects on late event related potential activity (P3s and SPs) during long foreperiod signaled reaction-time (RT) tasks (Podlesny and Dustman 1982). We found that negative SPs were larger for middle-aged and elderly subjects than for young adults, and that P3 amplitude at Pz was smaller for elderly than for younger subjects. The differences between young and elderly on both measures was in the direction of less positivity for the elderly. This result might also be interpreted to indicate reduced inhibitory activity in elderly subjects. In the present study, we recorded SP and P3 responses of young, middle-aged, and elderly adults during respond-withhold tasks to determine if the results would be consistent with the hypothesis that these ERPs may reflect reduced neural inhibition in the elderly. We also studied late activity (LA) following the P3 because late inhibitory activity might be expected to appear while subjects withhold their responses. LA corresponds in latency range to the 'slow wave' which has been reported to appear under similar conditions as the P3 (Sandquist et al. 1980) but may also be related to motor responses (McCarthy and Donchin 1976).
0013-4649/84/$03.00 ,,i: 1984 Elsevier Scientific Publishers Ireland, Ltd.
AGING. RESPOND-WITHHOLD TASKS, AND ERPs
Method
Subjccls There were 16 young, 16 middle-aged, and 16 elderly right-handed subjects (age ranges 26-35, 45 55, and 65 73 years; mean ages 30, 50 and 69 years). There were 11 women and 5 men in each group. All subjects were selected for good health and paid for participation.
Apparatus Each subject sat on a comfortable chair during data collection. The subject's eyes were 1.2 m from a Ball C R T monitor on which visual stimuli were presented. The RT switch was a spring-loaded toggle switch mounted at one end of a cylinder. The subject held the cylinder using the fingers of the right hand with the toggle switch at thumb level. Switch closure was by abduction of the thumb (movement of the thumb ventrally at a right angle from a plane through the palm of the hand). The switch was designed to require contraction of the thenar muscles so that electromyograms (EMGs) which reliably showed muscle activity during switch closure could be obtained. The EMGs were used to assure that no muscle activity occurred during withhold trials. A Terak Model 8510 computer was used for stimulus generation and on-line collection of electrophysiological and RT data. Electrophysiological recordings were made with a Grass Model 78B polygraph. Electroencephalograms were recorded with DC amplifiers from Fz, Cz, and Pz with a linked ear reference, using A g / A g C I electrodes (attached with collodion) and a NaCI electrolyte. To reduce DC drift, the electrodes were anodized regularly in a 5% saline solution and soaked in a 1% saline solution with leads shorted when not in use, Prior to each session, the EEG amplifiers were calibrated with an external sine wave signal at the gain and DC setting used for recording. No amplifier drift was present during any calibration. To avoid systematic error caused by possible differences in the amplifiers and analogue to digital converters contributing erroneously to electrode location effects, the recording channels for Fz, Cz, and Pz were varied regularly from subject to subject during
131
data collection. The number of subjects assigned to the different combinations of recording channels with electrode locations was approximately equal for the different age groups. Vertical electro-oculograms (VEOGs) were recorded with a DC amplifier from Ag/AgC1 Beckman miniature biopotential electrodes filled with NaCI electrolyte and attached to the skin over the supra- and suborbital ridges with adhesive collars. Electromyograms were recorded with a 100 msec time constant from 2 gold-plated cup electrodes filled with NaC1 electrolyte and attached with collodion about 2 cm apart on the skin over the longitudinal axis of the thenar eminence of the right hand. EEGs, EOGs, and E M G s were sampled at 1000 Hz. Successive values were averaged, giving 250 data points/sec for measurement of P3s and LA, 50 points/sec for SP measurement, and 20 points/sec for EOG measurement. Absolute E M G deviations from a mean baseline were averaged to give 50 data points/sec.
Procedure Persons responding to ads for paid subjects were screened by phone regarding their health. Those whose verbal responses indicated adequate health received a brief verbal description of the purposes of the experiment and the procedures to be used and were scheduled for a testing session. The session consisted of two parts: health screening, and testing on two respond-withhold RT tasks. An experimenter briefly questioned the subject and his/her medical history. Subjects indicating a history of cardiovascular or neurological dysfunction were excluded. Subjects were given a test of resting blood pressure (seated) by the auscultation method, and those having diastolic blood pressures more than 10 mm Hg above the values for healthy individuals in their age group were excluded. The age group norms given by Brobeck (1973) were used (the age 60 norms were used for subjects 60 years and older). Subjects were also tested for critical flicker fusion, but those data are not reported here. A Snellen chart was used to test vision, and a corrected visual acuity of 20/50 or better was required. Subjects thus selected were given a rest period prior to performing the respond-withhold tasks.
132 After the rest period the electrodes were attached. Each subject then performed a signaled respond-withhold task and a non-signaled respond-withhold task. Each task consisted of a block of 60 trials during which respond and withhold trials occurred randomly with a probability of 0.5. During the non-signaled task, the stimulus for each trial was the letter ' X ' or the letter 'O.' Subjects were instructed to press the switch immediately in response to one letter and not to press (to avoid thumb movement) in response to the other. During the signaled task, the letter was followed 3 sec later by a rectangle which was the imperative signal. Subjects were instructed to press the switch when they saw the rectangle if it was preceded by one letter and not to press if the rectangle was preceded by the other. Subjects were instructed to perform quickly and accurately. EMG was recorded to assure that subjects performed the respond-withhold tasks as instructed (i.e., that they did not produce EMG activity during withhold trials). Subjects were asked to fixate a small square (9 mm × 9 ram; visual arc, 26') on the CRT and to avoid blinking during trials. The fixation square remained on during the entire trial and stimuli were superimposed on it. The letters and rectangle were 5 cm high and subtended a visual arc of about 2.5 °. The duration of each stimulus was 50 msec. Intertrial intervals were determined manually and ranged between 5 and 20 sec depending on the stability of the recordings, ocular fixation, and the presence of eye blinks. The mean intertrial intervals for the non-signaled task were: young, 10.1; middle-aged, 11.5; and elderly, 10.6 sec. The corresponding values for the signaled task were 13.2, 14.4, and 13.8 sec, respectively. The 4 combinations of task order (signaled-non-signaled or non-signaled-signaled) with respond letter ('X' or ' O ' ) were repeated pseudo-randomly for each successive 4 subjects within each age group. The same order was used for each of the age groups. Data reduction Error trials and trials with EOG artifact were rejected. Examination of mean EOGs for the selected trials revealed that EOG artifact did not contribute substantially to the averaged EEG for
J.A. PODLESNY ET AL. any combination of conditions in the study. EMGs were averaged for the selected respond and withhold trials for each subject to determine if any subjects had produced muscular activity during withhold trials. While substantial EMG responses were recorded during respond trials for all subjects, none of the subjects produced substantial averaged activity indicative of covert muscle activity during withhold trials. Thus, trials selected for analysis were performed correctly and were free of artifacts. The mean number of trials averaged per subject for each trial type was 22 for the non-signaled task and 18 for the signaled task. Within each task, the mean number of trials averaged was approximately equal for all combinations of age, performance and trial type. The minimum number of trials averaged was 10. Each subject's RTs for respond trials and EEGs for respond and withhold trials were averaged. Measures of SP amplitude, P3 amplitude, P3 latency, and LA were obtained from the averaged data for each electrode location. For both kinds of trials, the baseline was the mean level of the averaged EEG during the 200 msec interval prior to the onset of letter presentation. SP amplitude was measured from the signaled task data and was the mean deviation from baseline for each of the four 500 msec intervals between 1 and 3 sec after the onset of the letter. P3 amplitude, P3 latency, and LA were measured from the non-signaled task data. The peak of the P3 was detected by scanning the averaged EEG data for the earliest maximum positive deflection from baseline between 248 and 500 msec after the onset of the letter. P3 amplitude was the difference between the mean level for the detected peak and baseline. P3 latency was the time in msec between the onset of the letter and the midpoint of the peak interval. LA was the difference from baseline for the mean level of the averaged EEG between 500 and 800 msec after the onset of the letter. EMGs were analyzed to assure that subjects did not produce muscle activity during withhold trials. Since no subject produced substantial EMG activity during withhold trials, and differences between EMGs to respond and withhold trials were attributable to instructions, EMGs were not analyzed further.
AGING, RESPOND-WITHHOLD TASKS, AND ERPs Mean differences for each measure were tested with analysis of variance (ANOVA). Between subjects factors were age (young, middle-aged, elderly) and performance (fast subjects, slow subjects). To obtain the performance subgroups, the subjects in each age group were divided according to the median RT during respond trials for the task from which the measure was obtained. The subjects assigned to the fast and slow groups did not exactly correspond for the two tasks. The number of subjects categorized differently for the two tasks was: young, 2; middle-aged, 8; and elderly, 4. Within subjects factors were electrode location (Fz, Cz, Pz) and trial type (respond, withhold). For the SP analysis, an additional within subjects factor was intervals (the four 500 msec intervals between 1 and 3 sec after onset of the letter). Further comparisons were made using t tests and Newman-Keuls tests. Relationships between measures were tested with Pearson r correlation coefficients.
Results
Reaction time Mean RT increased significantly with age for the non-signaled task, F (2, 42) = 12.44, P < 0.001, M.S.E. = 1632.90 (mean RTs for young, middleaged, and elderly subjects were 511, 545, and 582 msec, respectively). Although a similar age trend was observed for the signaled task, the group means (407, 451, and 479 msec, respectively) were not significantly different. Sustained potential Grand mean ERPs for age, electrode location and trial type during the 3 sec foreperiod of the signaled task are shown in Fig. 1. SP amplitude increased with successive 500 msec intervals between 1 sec after the letter and the onset of the rectangle for the combined ERPs from all subjects, electrode locations, and trial types, F (3, 126)= 18.36, P < 0.001, M.S.E. = 7.34. There were no significant interactions of age or performance with intervals. Although none of the mean SP amplitudes for withhold trials was positive, respond trials produced significantly more negative mean
133 SP amplitudes than withhold trials ( - 7 . 9 and - 3.3/~V, respectively, F (1, 42) = 59.84, P < 0.001, M.S.E. = 102.08). As shown in Fig. 2 and Table I, there was a significant age × electrode location x trial type interaction in SP amplitude for the combined intervals, F (4, 84)= 3.03, P < 0.05, M.S.E. = 12.30. Paired comparisons between the age group means for each electrode location and trial type showed that the elderly group had significantly larger negative amplitude at Cz than the middle-aged, q (3, 8 4 ) = 4.63, P < 0.01, and young, q (2, 84)= 4.26, P < 0.01, subjects during respond trials. None of the other comparisons were significant. The moment coefficient of skewness was computed for the Cz respond SP amplitudes of the elderly group to determine if the large negative mean amplitude might be due to a few extreme values. Skewness was slight (3q = 0.12), indicating there was a general shift toward more negative SP amplitude at Cz for the elderly group. As shown in Table II, there was a significant performance x electrode location × trial type interaction in SP amplitude, F (2, 84)= 4.70, P < 0.05, M.S.E. = 12.30. Comparisons of fast vs. slow subjects from the combined age groups for each electrode location and trial type showed that fast subjects had significantly larger mean SP ampli-
RESPOND
WITHHOLD
Kz ~ : ~f', ~ "~"'~~.~..: . ......... ...
V
.... •
I O
I 1
I 2
~i=: .~'~'~*~
Middle-aged
"
Elderly
I 3
j5~V +
I O
I 1
I 2
I 3
sec
Fig. l. Grand mean event-related potentials during the signaled task. The onset of the letter (ready stimulus) was at 0 sec and the onset of the imperative signal was at 3 sec.
134
J.A. PODLESNY
TABLE Mean
I
~-14.0
FZ
Cz
i'
s u s t a i n e d p o t e n t i a l a m p l i t u d e s (/.tV) for a g e x e l e c t r o d e
location x trial type. Each value is the average of four 500 msec i n t e r v a l m e a s u r e s b e t w e e n 1 a n d 3 sec after onset of t h e letter. Young
Middle-aged
Elderly
Z l - -10.0
~
o Fz
- 6.0
- 7.0
7.0
Cz
- 9.1
8.7
- 12.8
Pz
-6.8
-6.1
-7.4
OEIderly
r.
.
,ES~ORO
w,,..o~o
Pz
".
"o -aged "•"~LI, .... /
/
,~,.-
:E,.z, - 2.0
Respond trials
ET AL.
bE
r,
.
.ES~O.O
W....O~O
TRIAL TYPE
.E~;O.O W.,.'.OLO b
F i g . 2. M e a n s u s t a i n e d p o t e n t i a l a m p l i t u d e s of young, m i d d l e a g e d , a n d e l d e r l y subjects during respond and w i t h h o l d t r i a l s at
Withhold trials"
Fz, Cz, a n d Pz.
Fz
- 3.0
- 3.6
Cz
- 4.5
- 3.3
- 2.2 - 2.7
Pz
- 4.4
- 3.2
- 2.6
tude than slow subjects during respond trials at Cz, t ( 4 6 ) = 1.86, P < 0.05, S.E. = 1.61. N o n e of the other comparisons were significant.
P3 latency The grand mean ERPs for age, electrode location, and trial type during the 1000 msec following the letter in the non-signaled task are shown in Fig. 3. Both the respond and withhold conditions produced prominent grand mean P3 responses at all electrode locations in all age groups. Mean P3 latency was similar for young (362 msec) and middle-aged (360 msec) subjects, but significantly longer for elderly subjects (401 msec), F (2, 42) = 6.07, P < 0.01, M.S.E. = 8531.24. P3 latency for respond trials (364 msec) was significantly shorter
TABLE Mean
than P3 latency for withhold trials (385 msec), F (1, 42) = 11.17, P < 0.01, M.S.E. = 2701.14. However, there were no significant age x trial type interactions.
P3 amplitude As shown in Fig. 4 and Table III, there was a significant age × electrode location × trial type interaction in P3 amplitude, F (4, 8 4 ) = 3.83, P < 0.01, M.S.E. = 3.19. During respond trials, mean P3 amplitude decreased significantly with increasing age at Cz and at Pz (see Table III). However, at Fz elderly subjects as well as young subjects had significantly larger mean P3 amplitude than middle-aged subjects. Mean P3 amplitudes at Fz for young and
RESPOND
.....
II
WITHHOLD
'.-.'
~
s u s t a i n e d p o t e n t i a l a m p l i t u d e s (p.V) for performancex
e l e c t r o d e l o c a t i o n x t r i a l type.
Fast subjects
j:d ~
.~il
Slow subjects
Respond trials Fz
- 6.9
- 6.5
....
Cz
- 11.7
- 8.7
....... E l d e r l y
Pz
- 7.4
- 6.2
o"
Withhold trials Fz
- 3.1
- 2.8
Cz
- 3.4
- 3.6
Pz
- 3.5
- 3.3
'40"
;o'oo
Middle-aged
~7
_ !5pV
o'"
"soo mseg
"1ooo
F i g . 3. Grand mean e v e n t - r e l a t e d p o t e n t i a l s d u r i n g t h e non-signaled task. T h e onset of t h e l e t t e r ( i m p e r a t i v e stimulus) was a t 0 msec.
A G I N G , R E S P O N D - W I T H H O L D TASKS, A N D ERPs il 21.o~
RESPOND TRIALS
19.0~
"-..
g ,,.o1.,-
"~,,
.......j
. "-,
xs.o
L II
_.13.0 < 1
subjects had significantly larger mean P3 amplitudes than middle-aged subjects at Cz and Pz. At Fz elderly subjects produced larger mean P3 amplitude than either of the younger groups, while young subjects had larger mean P3 amplitudes than middle-aged subjects.
WITHHOLD TRIALS
".%<
o.
a
~
135
" %., %
%
,."
.
; .01.
....,,..o
Fz o,e.o Cz
% YOUNG MIDDLE ELDERLY -AGED
ez
,~..-o
•
1
1
YOUNG MIDDLE ELDERLY -AGED
Fig. 4. Mean P3 amplitudes of young, middle-aged, and elderly subjects during respond and withhold trials at Fz, Cz, and Pz.
elderly subjects did not differ significantly during respond trials. During withhold trials, young subjects had larger mean P3 amplitudes than middle-aged or elderly subjects at Cz and at Pz. However, elderly
Late activity Mean late activity, which followed the period for measurement of the P3, was positive and significantly larger during withhold trials (4.4 ffV) than during respond trials (0.3 ffV) over all subjects and electrode locations, F (1, 42)= 25.30, P < 0.001, M.S.E. = 47.40. The age x trial type interaction was significant, F (2, 42) = 7.06, P < 0.01, M.S.E. = 47.40, indicating that elderly subjects had a larger difference in late activity between respond and withhold trials than the younger groups (elderly, - 1 . 4 and 6.9 ffV; middle-aged, 2.0 and
T A B L E Ill Mean P3 amplitude (ffV) for age x electrode location x trial type and significant mean comparisons among young (Y). middle-aged (M), and elderly (E) groups. Statistics which indicate significance of the mean comparisons are shown below. Young
Middle-aged
Elderly
Mean comparisons
16.9 17.9 20.4
14.8 14.7 18.6
17.4 11.3 15.0
Y, E > M * * Y > M,E**; M > E** Y > M,E**; M > E**
16.1 18.6 19.0
14.4 13.5 14.5
17.9 15.4 15.9
E > Y,M**; Y > M * * Y > E,M**; E > M * * Y > E,M**; E > M*
Respond trials Fz Cz Pz
Withhold trials Fz Cz Pz
Young vs. middle-aged
r,~
Young vs. elderly
q
Middle-aged vs. elderly
r,~
q
r,~
q
Respond trials" Fz Cz Pz
2, 84 2, 84 2, 84
4.63 7.12 4.20
3, 84 3, 84
14.67 12.15
3, 84 2, 84 2, 84
5.83 7.55 7.95
2, 84 3, 84 3, 84
3.83 11.32 9.96
2, 84 2, 84 2, 84
4.13 7.17 6.75
3, 84 2, 84 2, 84
7.96 4.15 3.21
Withhold trials Fz Cz Pz * P < 0.05; ** P < 0.01.
136
J.A. P O D L E S N Y Fz
cz
vz
~. - 4 . 0
r. F--
•
".
0.0
~+4o
11
o..
÷s0
;~'o RESPOND
WITHHOLD
RESPOND TRIAL
WITHHOLD TYPE
RESPOND
I
WITHHOLD
Fig. 5. Mean late activity of young, middle-aged, and elderly subjects during respond and withhold trials at Fz, Cz, a n d Pz.
3.6/~V; young, 0.4 and 2.7/~V, respectively). The age × electrode location × trial type interaction was also significant, F (4, 84)= 5.84, P < 0.001, M.S.E. = 4.81. As shown in Fig. 5 and Table IV, the pattern of mean late activity for the different age groups and trial types varied across electrode locations. During respond trials at Fz,
ET AL.
young subjects had a larger negative mean than the middle-aged and elderly subjects, while the middle-aged and elderly subjects did not differ significantly. During withhold trials at Fz, mean late activity showed significantly larger positive values with increasing age (see Fig. 5 and Table IV). At Cz, elderly subjects produced the largest negative mean late activity during respond trials and the largest mean positive late activity during withhold trials. The Cz means for the young and middle-aged groups differed significantly for respond trials but not for withhold trials. All of the means were positive at Pz. The means for the young and middle-aged groups did not differ significantly for either respond or withhold trials. However, at Pz the elderly group had significantly smaller positive mean late activity during respond trials and significantly larger positive mean late activity during withhold trials.
T A B L E IV
Mean late activity (p,V) for age × electrode location x trial type and significant mean comparisons among young (Y), m i d d l e - a g e d (M), and elderly (E) groups. Statistics which indicate significance of the mean comparisons are shown below. Young
Middle-aged
Elderly
Mean comparisons
- 3.3 -1.9 6.4
- 1.3 0.3 7.0
- 1.2 -7.5 4.4
E,M > Y * Y,M > E**: M > Y** M > E * * : Y > E*
0.8 2.0 5.4
2.9 3.0 4.9
7.4 5.9 7.5
E > Y,M**: M > Y** E > Y,M** E > Y,M**
Respondtrials Fz Cz Pz
WithhoM trials" Fz Cz Pz
Young vs. middle-aged
Young vs. elderly
Middle-aged vs. elderly
r,~
q
r,~
q
r,~
q
2, 84 2, 84
3.66 3.95 -
3, 84 2, 84 2, 84
3.68 10.22 3.65
3, 84 3, 84
14.18 4.85
2, 84
3.97 -
3, 84 3, 84 2, 84
12.05 7.18 3.84
2, 84 2, 84 3, 84
8.09 5.29 4.79
Respondtrials Fz Cz Pz
Withhold trials Fz Cz Pz * P < 0.05; ** P < 0.01.
-
AGING, RESPOND-WITHHOLD TASKS, AND ERPs
f
ELDERLY SUBJECTS, Cz
7
Respond trials ....... Withhold trial,
.: / •
0
T.
-42 -37 -32 -27 -22 -17 -12 -7 -2 3 8 113 18 MEAN LATE ACTIVITY (INTERVAL MIDPOINTS,pV)
Fig. 6. Frequency distributions comparing the mean amplitude of late activity for elderly subjects during respond and withhold trials at ('z.
To determine if the unexpectedly large negative respond and positive withhold LA means of elderly subjects at Cz were due to a few extreme values, the moment coefficients of skewness were computed for these distributions. Both distributions were moderately negatively skewed (respond trials, y~ = -0.83; withhold trials, Yt = -1.33). As may be seen in Fig. 6, the distribution for respond trials contained predominantly negative values, and the distribution for withhold trials contained predominantly positive values. Correlations between SP and LA
Since the large negative SP and LA values at Cz during respond trials in the elderly group suggested a common process, we computed correlation coefficients between SP amplitude and LA for respond trials at Cz for the different age groups. SP amplitude was significantly related to LA in the elderly group ( r = 0.51, P < 0.05) but not in the young and middle-aged groups (rs = 0.08 and 0.31, respectively).
Discussion
The hypothesis that inhibitory activity might produce more positive potentials during withhold trials was generally supported by the results with LA but not by those with P3 amplitude. LA at Fz and Cz shifted from values which were generally negative during respond trials to values which
137
were positive during withhold trials. Surprisingly, this effect was strongest with elderly subjects. However, there was no definite trend toward larger P3 amplitude during withhold trials. This finding with P3 amplitude is contrary to that of Karlin et al. (1970). The results with SP amplitude are equivocal with respect to this hypothesis, since the means were negative for withhold trials, as well as respond trials. The fact that withhold trials were significantly less negative than respond trials might be interpreted to indicate a summative positive component or simply less negative activity for trials during which the subjects were not required to respond. Certain of the results may be interpreted as indicating weakened inhibitory activity in the elderly. The elderly group produced the largest negative SPs and LA at Cz during respond trials and this might be explained by stronger excitation due to weakened inhibition. However, the elderly group also produced the most positive LA during withhold trials at all electrode sites. If it is assumed that positive potentials reflect inhibitory activity, the results with LA suggest stronger inhibitory activity in the elderly subjects. Alternatively, the large SPs and LA in the elderly group could indicate weakened inhibitory control of processes which give rise to these potentials. There were several puzzling results. Elderly subjects produced the largest negative SPs at Cz during respond trials. This agrees with our previous finding in a different sample of subjects (Podlesny and Dustman 1982) and appears paradoxical since in both studies, faster performing subjects produced larger negative SPs while the elderly groups performed most slowly. The large negative LA which appeared at Cz during respond trials in the elderly group was unexpected and was related to the large negative SPs in that group. The large positive LA during withhold trials in the elderly subjects is also surprising. Finally, the reduction of P3 amplitude with age at Pz and Cz parallels previous findings (Tecce et al. 1980, 1982; Podlesny and Dustman 1982). Tecce and coworkers have also noted that frontal P3 amplitude was not reduced in elderly people, and they suggested that this may indicate strength of orienting a n d / o r dishabituation processes in the elderly (Tecce et al.
138 1982). The present study produced larger frontal P3 amplitude in the elderly group, particularly during withhold trials. The shift in maximal P3 amplitude from Cz and Pz in the young group to Fz in the elderly group suggests a more frontal distribution of function in the elderly. The present results are subject to a variety of interpretations of which a few might be suggested. Aside from the question of whether inhibitory processes are reflected in these late potentials, it is noteworthy that the elderly people produced significantly larger negative SPs at Cz during respond trials, significantly larger negative and positive LA at Cz during respond and withhold trials, respectively, and significantly larger P3 amplitudes at Fz during withhold trials relative to the younger groups. These findings suggest the possibility that the elderly subjects may have applied greater effort to performance on the tasks. Perhaps adequate performance of the tasks was more challenging for the elderly and required them to mobilize neural resources to a greater extent. Such mobilization of (as yet indefinite) processes could be the source of their relatively large potentials. It may be suggested that more difficult respond-withhold tasks might also enhance these potentials in younger subjects. Since all subjects were selected for good health, these results may also be interpreted as arising from possible differential effects of the stringent selection criteria for our subjects. The elderly group may represent 'survivors,' individuals who have achieved old age in good health (Siegler et al. 1980). Large potentials among such healthy elderly subjects might reflect adaptive mechanisms which may have been innate or achieved developmentally, possibly in response to the decline of other functions.
Summary Differences in late event-related potentials (ERPs) during active respond-withhold tasks were tested with 16 young, 16 middle-aged, and 16 elderly subjects (mean ages were 30, 50 and 69 years, respectively). All subjects were selected for good health. While monopolar electroencephalograms were recorded from Fz, Cz, and Pz, subjects
J.A. PODLESNY ET AL. performed a signaled respond-withhold task to allow measurement of sustained potentials (SPs) and a non-signaled respond-withhold task for measurement of P3 responses and late activity (LA) which follows the P3. Respond and withhold trials were presented randomly with a probability of 0.5 in both tasks. Compared with younger subjects, the elderly group produced significantly larger negative mean SP amplitude during respond trials at Cz, significantly larger negative mean LA during respond trials at Cz, and significantly larger positive LA during withhold trials at all electrode locations. The elderly group also produced significantly larger P3 amplitude at Fz. The results are discussed in relation to the hypothesis that inhibition may be weaker in elderly subjects. Two further possible interpretations of the larger ERPs in elderly subjects are suggested: (1) the elderly group may have been more challenged by the tasks and their larger ERPs may reflect greater neural effort; a n d / o r (2) the health selection procedures may have produced a group of elderly 'survivors' whose large ERPs may be characteristic of robust individuals or may reflect adaptations to aging-related changes in neural function.
Rdsum6 Vieillissement et thche d'arrOt de r~ponse." effets sur les potentiels permanents, les rkponses P3 et I'activitb tardive
On a recherch6 des diff6rences entre potentiels tardifs li6s ~ l'6v6nement d6velopp6s au cours de t~che active de r6ponse ou d'abstention de r6ponse, chez 16 jeunes, 16 sujets d'~ge moyen et 16 sujets plus ~g6s (ages moyens 30, 50 et 69 ans respectivement). Tousles sujets s61ectionn6s 6taient en bonne sant6. Pendant que les sujets effectuaient une telle thche signal6e, un EEG monopolaire 6tait recueilli en Fz, Cz et Pz et l'on mesurait le potentiel stable (PS); une thche analogue, non signal6e, servait 6valuer les r6ponses P3 et l'activit6 tardive subs6quente. Les essais avec r6ponse ou non-r6ponse 6taient pr6sent6s al6atoirement, avec pour chacun une probabilit6 de 0,5.
AGING, RESPOND-WITHHOLD TASKS. AND ERPs
Par rapport h des sujets plus jeunes, les sujets ~g6s ont d6velopp6 une activit6 tardive moyenne n6gative significativement plus grande pendant les essais avec r6ponse en Cz, de m6me qu'une activit6 tardive positive plus grande pendant les essais ou la r6ponse 6tait emp6ch6e, ceci sur toutes les 61ectrodes. Le groupe le plus fi.g6 a 6galement pr6sent6 une P3 d'amplitude plus grande h Fz. Ces r6sultats sont discut6s en relation avec l'hypoth6se de la possibilit6 d'une inhibition plus faible chez les sujets ~tg6s. Deux autres interpr6tations possibles sont sugg6r6es: (1) le groupe ~g6 s'est peut 6tre plus investi dans les tests et l'ampleur de leurs potentiels pourrait refl6ter un plus grand effort c6r6bral; et/ou (2) le crit6re de s61ection par la sant6 a produit un groupe ag6 de 'survivants' pour lesquels l'ampleur des potentiels 6voqu6s pourrait 6tre caract6ristique d'individus robustes ou pourrait refl6ter des adaptations h des modifications de la fonction neuronale li6es au vieillissement. The technical assistance provided by Albert Johnson and Dr. Cameron Schlehuber is gratefully acknowledged.
References Adolfsson, R., Gottfries, C.G., Roos, B.E. and Winblad, B. Post mortem distribution of dopamine and homovanillic acid in human brain, variations related to age, and a review of the literature. J. neural Transm., 1979, 45: 81-105. Agrawal, H.C. and Himwich, W.A. Amino acids, proteins and monoamines of developing brain. In: W.A. Himwich (Ed.), Developmental Neurobiology. Thomas, Springfield, IL, 1970: 287-310. Beck, C.H.M. Functional implications of changes in the senescent brain: a review. Canad, J. neurol. Sci., 1978, 5: 417-424. Brobeck, J.R. Best and Taylor's Physiological Basis of Medical Practice, 9th edn. Williams and Wilkins, Baltimore, MD, 1973. Dustman, R.E. and Beck, E.C. The effects of maturation and aging on the waveform of visually evoked potentials. Electroenceph, clin. Neurophysiol., 1969, 26: 2-11. Dustman, R.E. and Snyder, E.W. Life-span changes in visually evoked potentials at central scalp. Neurobiol. Aging, 1981, 2: 303-308. Dustman, R.E., Snyder, E.W. and Schlehuber, C.J. Life-span alterations in visually evoked potentials and inhibitory function. Neurobiol. Aging, 1981, 2: 187-192. Frolkis, V.V. and Bezrukov, V.V. Aging of the Central Nervous System. Interdisciplinary Topics in Gerontology, Vol. 16. Karger, New York, 1979:131 pp.
139 Howard, R., Fenton, G.W. and Fenwick, P.B. Slow cerebral potentials in a 'go-no go' avoidance situation: a study on special hospital patients. In: C. Barber (Ed.), Evoked Potentials. University Park Press, Baltimore, MD, 1980: 515-523. Jakubczak, L.F. Age and animal behavior. In: C. Eisdorfer and M.P. Lawton (Eds.), The Psychology of Adult Development and Aging. American Psychological Association, Washington, DC, 1973: 98-111. Karlin, L., Martz, M.J. and Mordkoff, A.M. Motor performance and sensory-evoked potentials. Electroenceph. clin. Neurophysiol., 1970, 28: 307-313. McCarthy, G. and Donchin, E. The effect of temporal and event uncertainty in determining the waveforms of the auditory event related potential (ERP). Psychophysiology, 1976, 13: 581-590. McGeer, P.L. and McGeer, E.G. Chemistry of mood and emotion. Ann. Rev. Psychol., 1980, 31: 273-307. Podlesny, J.A. and Dustman, R.E. Age effects on heart rate, sustained potential, and P3 responses during reaction-time tasks. Neurobiol. Aging, 1982, 3: 1-9. Pradhan, S.N. Minireview: central neurotransmitters and aging. Life Sci., 1980, 26: 1643-1656. Rohrbaugh, J.W., Syndulko, K. and Lindsley, D.B. Brain wave components of the contingent negative variation in humans. Science, 1976, 191: 1055-1057. Rohrbaugh, J.W., Syndulko, K., Sandquist, T.F. and Lindsley, D.B. Synthesis of the contingent negative variation brain potential from noncontingent stimulus and motor elements. Science, 1980, 208: 1165-1168. Sandquist, T.F., Rohrbaugh, J.W., Syndulko, K. and Lindsley, D.B. Electrocortical signs of levels of processing: perceptual analysis and recognition memory. Psychophysiology, 1980, 17: 568-576. Scheibel, M.G. and Scheibel, A.B. Structural changes in the aging brain. In: H. Brody, D. Harmon and J.M. Ordy (Eds.), Aging, Vol. 1. Raven Press, New York, 1975: 11-37. Siegler, I.C., Nowlin, J.B. and Blumenthal, J.A. Health and behavior: methodological considerations for adult development and aging. In: L.W. Poon (Ed.), Aging in the 1980s. American Psychological Association, Washington, DC, 1980: 599-612. Simson, R., Vaughn, Jr., H.G. and Ritter, W. The scalp topography of potentials in auditory and visual g o / n o go tasks. Electroenceph. clin. Neurophysiol., 1977, 43: 864-875. Straumanis, J.J., Shagass, C. and Overton, D.A. Somatosensory evoked responses in Down's syndrome. Arch. gen. Psychiat., 1973, 29: 544-549. Tecce, J.J., Yrchik, D.A., Meinbresse, D., Dessonville, C.L. and Cole, J.O. CNV rebound and aging. I. Attention functions. In: H.H. Kornhuber and L. Deecke (Eds.), Progress in Brain Research. Vol. 54. Motivation, Motor and Sensory Processes of the Brain: Electrical Potentials, Behavior and Clinical Use. Elsevier/North-Holland Biomedical Press, Amsterdam, 1980: 552-561. Tecce, J.J., Cattanach, L., Yrchik, D.A., Meinbresse, D. and Dessonville, C.L. CNV rebound and aging. Electroenceph. clin. Neurophysiol., 1982, 54: 175-186.
130
Experimental Section AGING AND RESPOND-WITHHOLD TASKS: EFFECTS ON SUSTAINED POTENTIALS, P3 RESPONSES AND LATE ACTIVITY t J.A. PODLESNY 2, R.E. D U S T M A N 3 and D.E. S H E A R E R
Neuropsychology Research Laboratory', Veterans Administration Medical Center, Salt Lake Ci(v, UT 8414~ ¢U.S.A.) (Accepted for publication: January 11, 1984)
Reported effects on sustained potential 4 (SP, or CNV) and P3 responses of young adults during active respond-withhold tasks have been interpreted as consistent with the hypothesis that inhibitory cortical activity may be reflected by greater positive late potentials (Howard et al. 1980). These tasks require subjects to produce motor responses during some trials and withhold responses during other randomly presented trials. Presumably, since subjects must inhibit motor activity during withhold trials, their ERPs may reflect neural inhibition. P3 amplitude was reportedly larger during withhold trials than during respond trials (Karlin et al. 1970; Simson et al. 1977). Elderly people are believed to have weakened inhibitory function (Jakubczak 1973; Scheibel and Scheibel 1975; Beck 1978; Frolkis and Bezrukov 1979; McGeer and McGeer 1980), possibly related to reduced neurotransmitter activity in late life
1 This research was supported by the Medical Research Service of the U.S. Veterans Administration and by the U.S. National Institutes of Health Grants AG05126 and AG00568. 2 J.A. Podlesny is currently with the Laboratory Division, Federal Bureau of Investigation, Washington, DC 20535, U.S.A. 3 Requests for reprints should be addressed to Dr. Robert E. Dustman, Neuropsychology Research Laboratory (151A), Veterans Administration Medical Center, Salt Lake City, UT 84148, U.S.A. 4 We have used the term 'sustained potential' rather than 'contingent negative variation' because studies of functional and topographic differences in the early and late portions of the wave form (Rohrbaugh et al. 1976, 1980) suggest 'that it is constituted largely of separate noncontingent elements related to sensory and motor processes' (Rohrbaugh et al. 1980, p. 1165). 'Sustained potential' is a relatively neutral term which describes the sustained quality of the observed potentials without suggesting contingency which is less obvious.
(Agrawal and Himwich 1970; Beck 1978; Adolfsson et al. 1979; McGeer and McGeer 1980; Pradhart 1980). We and others have reported that evoked potentials recorded from elderly subjects in a passive state (no task involved) often show characteristics which suggest reduced inhibitory function, i.e., enhanced wave amplitude (Dustman and Beck 1969; Straumanis et al. t973; Dustman and Snyder 1981) and a relative inability of the visual system to differentiate patterned from unpatterned flashes (Dustman et al. 1981). More recently, we studied aging effects on late event related potential activity (P3s and SPs) during long foreperiod signaled reaction-time (RT) tasks (Podlesny and Dustman 1982). We found that negative SPs were larger for middle-aged and elderly subjects than for young adults, and that P3 amplitude at Pz was smaller for elderly than for younger subjects. The differences between young and elderly on both measures was in the direction of less positivity for the elderly. This result might also be interpreted to indicate reduced inhibitory activity in elderly subjects. In the present study, we recorded SP and P3 responses of young, middle-aged, and elderly adults during respond-withhold tasks to determine if the results would be consistent with the hypothesis that these ERPs may reflect reduced neural inhibition in the elderly. We also studied late activity (LA) following the P3 because late inhibitory activity might be expected to appear while subjects withhold their responses. LA corresponds in latency range to the 'slow wave' which has been reported to appear under similar conditions as the P3 (Sandquist et al. 1980) but may also be related to motor responses (McCarthy and Donchin 1976).
0013-4649/84/$03.00 ,,i: 1984 Elsevier Scientific Publishers Ireland, Ltd.
AGING. RESPOND-WITHHOLD TASKS, AND ERPs
Method
Subjccls There were 16 young, 16 middle-aged, and 16 elderly right-handed subjects (age ranges 26-35, 45 55, and 65 73 years; mean ages 30, 50 and 69 years). There were 11 women and 5 men in each group. All subjects were selected for good health and paid for participation.
Apparatus Each subject sat on a comfortable chair during data collection. The subject's eyes were 1.2 m from a Ball C R T monitor on which visual stimuli were presented. The RT switch was a spring-loaded toggle switch mounted at one end of a cylinder. The subject held the cylinder using the fingers of the right hand with the toggle switch at thumb level. Switch closure was by abduction of the thumb (movement of the thumb ventrally at a right angle from a plane through the palm of the hand). The switch was designed to require contraction of the thenar muscles so that electromyograms (EMGs) which reliably showed muscle activity during switch closure could be obtained. The EMGs were used to assure that no muscle activity occurred during withhold trials. A Terak Model 8510 computer was used for stimulus generation and on-line collection of electrophysiological and RT data. Electrophysiological recordings were made with a Grass Model 78B polygraph. Electroencephalograms were recorded with DC amplifiers from Fz, Cz, and Pz with a linked ear reference, using A g / A g C I electrodes (attached with collodion) and a NaCI electrolyte. To reduce DC drift, the electrodes were anodized regularly in a 5% saline solution and soaked in a 1% saline solution with leads shorted when not in use, Prior to each session, the EEG amplifiers were calibrated with an external sine wave signal at the gain and DC setting used for recording. No amplifier drift was present during any calibration. To avoid systematic error caused by possible differences in the amplifiers and analogue to digital converters contributing erroneously to electrode location effects, the recording channels for Fz, Cz, and Pz were varied regularly from subject to subject during
131
data collection. The number of subjects assigned to the different combinations of recording channels with electrode locations was approximately equal for the different age groups. Vertical electro-oculograms (VEOGs) were recorded with a DC amplifier from Ag/AgC1 Beckman miniature biopotential electrodes filled with NaCI electrolyte and attached to the skin over the supra- and suborbital ridges with adhesive collars. Electromyograms were recorded with a 100 msec time constant from 2 gold-plated cup electrodes filled with NaC1 electrolyte and attached with collodion about 2 cm apart on the skin over the longitudinal axis of the thenar eminence of the right hand. EEGs, EOGs, and E M G s were sampled at 1000 Hz. Successive values were averaged, giving 250 data points/sec for measurement of P3s and LA, 50 points/sec for SP measurement, and 20 points/sec for EOG measurement. Absolute E M G deviations from a mean baseline were averaged to give 50 data points/sec.
Procedure Persons responding to ads for paid subjects were screened by phone regarding their health. Those whose verbal responses indicated adequate health received a brief verbal description of the purposes of the experiment and the procedures to be used and were scheduled for a testing session. The session consisted of two parts: health screening, and testing on two respond-withhold RT tasks. An experimenter briefly questioned the subject and his/her medical history. Subjects indicating a history of cardiovascular or neurological dysfunction were excluded. Subjects were given a test of resting blood pressure (seated) by the auscultation method, and those having diastolic blood pressures more than 10 mm Hg above the values for healthy individuals in their age group were excluded. The age group norms given by Brobeck (1973) were used (the age 60 norms were used for subjects 60 years and older). Subjects were also tested for critical flicker fusion, but those data are not reported here. A Snellen chart was used to test vision, and a corrected visual acuity of 20/50 or better was required. Subjects thus selected were given a rest period prior to performing the respond-withhold tasks.
132 After the rest period the electrodes were attached. Each subject then performed a signaled respond-withhold task and a non-signaled respond-withhold task. Each task consisted of a block of 60 trials during which respond and withhold trials occurred randomly with a probability of 0.5. During the non-signaled task, the stimulus for each trial was the letter ' X ' or the letter 'O.' Subjects were instructed to press the switch immediately in response to one letter and not to press (to avoid thumb movement) in response to the other. During the signaled task, the letter was followed 3 sec later by a rectangle which was the imperative signal. Subjects were instructed to press the switch when they saw the rectangle if it was preceded by one letter and not to press if the rectangle was preceded by the other. Subjects were instructed to perform quickly and accurately. EMG was recorded to assure that subjects performed the respond-withhold tasks as instructed (i.e., that they did not produce EMG activity during withhold trials). Subjects were asked to fixate a small square (9 mm × 9 ram; visual arc, 26') on the CRT and to avoid blinking during trials. The fixation square remained on during the entire trial and stimuli were superimposed on it. The letters and rectangle were 5 cm high and subtended a visual arc of about 2.5 °. The duration of each stimulus was 50 msec. Intertrial intervals were determined manually and ranged between 5 and 20 sec depending on the stability of the recordings, ocular fixation, and the presence of eye blinks. The mean intertrial intervals for the non-signaled task were: young, 10.1; middle-aged, 11.5; and elderly, 10.6 sec. The corresponding values for the signaled task were 13.2, 14.4, and 13.8 sec, respectively. The 4 combinations of task order (signaled-non-signaled or non-signaled-signaled) with respond letter ('X' or ' O ' ) were repeated pseudo-randomly for each successive 4 subjects within each age group. The same order was used for each of the age groups. Data reduction Error trials and trials with EOG artifact were rejected. Examination of mean EOGs for the selected trials revealed that EOG artifact did not contribute substantially to the averaged EEG for
J.A. PODLESNY ET AL. any combination of conditions in the study. EMGs were averaged for the selected respond and withhold trials for each subject to determine if any subjects had produced muscular activity during withhold trials. While substantial EMG responses were recorded during respond trials for all subjects, none of the subjects produced substantial averaged activity indicative of covert muscle activity during withhold trials. Thus, trials selected for analysis were performed correctly and were free of artifacts. The mean number of trials averaged per subject for each trial type was 22 for the non-signaled task and 18 for the signaled task. Within each task, the mean number of trials averaged was approximately equal for all combinations of age, performance and trial type. The minimum number of trials averaged was 10. Each subject's RTs for respond trials and EEGs for respond and withhold trials were averaged. Measures of SP amplitude, P3 amplitude, P3 latency, and LA were obtained from the averaged data for each electrode location. For both kinds of trials, the baseline was the mean level of the averaged EEG during the 200 msec interval prior to the onset of letter presentation. SP amplitude was measured from the signaled task data and was the mean deviation from baseline for each of the four 500 msec intervals between 1 and 3 sec after the onset of the letter. P3 amplitude, P3 latency, and LA were measured from the non-signaled task data. The peak of the P3 was detected by scanning the averaged EEG data for the earliest maximum positive deflection from baseline between 248 and 500 msec after the onset of the letter. P3 amplitude was the difference between the mean level for the detected peak and baseline. P3 latency was the time in msec between the onset of the letter and the midpoint of the peak interval. LA was the difference from baseline for the mean level of the averaged EEG between 500 and 800 msec after the onset of the letter. EMGs were analyzed to assure that subjects did not produce muscle activity during withhold trials. Since no subject produced substantial EMG activity during withhold trials, and differences between EMGs to respond and withhold trials were attributable to instructions, EMGs were not analyzed further.
AGING, RESPOND-WITHHOLD TASKS, AND ERPs Mean differences for each measure were tested with analysis of variance (ANOVA). Between subjects factors were age (young, middle-aged, elderly) and performance (fast subjects, slow subjects). To obtain the performance subgroups, the subjects in each age group were divided according to the median RT during respond trials for the task from which the measure was obtained. The subjects assigned to the fast and slow groups did not exactly correspond for the two tasks. The number of subjects categorized differently for the two tasks was: young, 2; middle-aged, 8; and elderly, 4. Within subjects factors were electrode location (Fz, Cz, Pz) and trial type (respond, withhold). For the SP analysis, an additional within subjects factor was intervals (the four 500 msec intervals between 1 and 3 sec after onset of the letter). Further comparisons were made using t tests and Newman-Keuls tests. Relationships between measures were tested with Pearson r correlation coefficients.
Results
Reaction time Mean RT increased significantly with age for the non-signaled task, F (2, 42) = 12.44, P < 0.001, M.S.E. = 1632.90 (mean RTs for young, middleaged, and elderly subjects were 511, 545, and 582 msec, respectively). Although a similar age trend was observed for the signaled task, the group means (407, 451, and 479 msec, respectively) were not significantly different. Sustained potential Grand mean ERPs for age, electrode location and trial type during the 3 sec foreperiod of the signaled task are shown in Fig. 1. SP amplitude increased with successive 500 msec intervals between 1 sec after the letter and the onset of the rectangle for the combined ERPs from all subjects, electrode locations, and trial types, F (3, 126)= 18.36, P < 0.001, M.S.E. = 7.34. There were no significant interactions of age or performance with intervals. Although none of the mean SP amplitudes for withhold trials was positive, respond trials produced significantly more negative mean
133 SP amplitudes than withhold trials ( - 7 . 9 and - 3.3/~V, respectively, F (1, 42) = 59.84, P < 0.001, M.S.E. = 102.08). As shown in Fig. 2 and Table I, there was a significant age × electrode location x trial type interaction in SP amplitude for the combined intervals, F (4, 84)= 3.03, P < 0.05, M.S.E. = 12.30. Paired comparisons between the age group means for each electrode location and trial type showed that the elderly group had significantly larger negative amplitude at Cz than the middle-aged, q (3, 8 4 ) = 4.63, P < 0.01, and young, q (2, 84)= 4.26, P < 0.01, subjects during respond trials. None of the other comparisons were significant. The moment coefficient of skewness was computed for the Cz respond SP amplitudes of the elderly group to determine if the large negative mean amplitude might be due to a few extreme values. Skewness was slight (3q = 0.12), indicating there was a general shift toward more negative SP amplitude at Cz for the elderly group. As shown in Table II, there was a significant performance x electrode location × trial type interaction in SP amplitude, F (2, 84)= 4.70, P < 0.05, M.S.E. = 12.30. Comparisons of fast vs. slow subjects from the combined age groups for each electrode location and trial type showed that fast subjects had significantly larger mean SP ampli-
RESPOND
WITHHOLD
Kz ~ : ~f', ~ "~"'~~.~..: . ......... ...
V
.... •
I O
I 1
I 2
~i=: .~'~'~*~
Middle-aged
"
Elderly
I 3
j5~V +
I O
I 1
I 2
I 3
sec
Fig. l. Grand mean event-related potentials during the signaled task. The onset of the letter (ready stimulus) was at 0 sec and the onset of the imperative signal was at 3 sec.
134
J.A. PODLESNY
TABLE Mean
I
~-14.0
FZ
Cz
i'
s u s t a i n e d p o t e n t i a l a m p l i t u d e s (/.tV) for a g e x e l e c t r o d e
location x trial type. Each value is the average of four 500 msec i n t e r v a l m e a s u r e s b e t w e e n 1 a n d 3 sec after onset of t h e letter. Young
Middle-aged
Elderly
Z l - -10.0
~
o Fz
- 6.0
- 7.0
7.0
Cz
- 9.1
8.7
- 12.8
Pz
-6.8
-6.1
-7.4
OEIderly
r.
.
,ES~ORO
w,,..o~o
Pz
".
"o -aged "•"~LI, .... /
/
,~,.-
:E,.z, - 2.0
Respond trials
ET AL.
bE
r,
.
.ES~O.O
W....O~O
TRIAL TYPE
.E~;O.O W.,.'.OLO b
F i g . 2. M e a n s u s t a i n e d p o t e n t i a l a m p l i t u d e s of young, m i d d l e a g e d , a n d e l d e r l y subjects during respond and w i t h h o l d t r i a l s at
Withhold trials"
Fz, Cz, a n d Pz.
Fz
- 3.0
- 3.6
Cz
- 4.5
- 3.3
- 2.2 - 2.7
Pz
- 4.4
- 3.2
- 2.6
tude than slow subjects during respond trials at Cz, t ( 4 6 ) = 1.86, P < 0.05, S.E. = 1.61. N o n e of the other comparisons were significant.
P3 latency The grand mean ERPs for age, electrode location, and trial type during the 1000 msec following the letter in the non-signaled task are shown in Fig. 3. Both the respond and withhold conditions produced prominent grand mean P3 responses at all electrode locations in all age groups. Mean P3 latency was similar for young (362 msec) and middle-aged (360 msec) subjects, but significantly longer for elderly subjects (401 msec), F (2, 42) = 6.07, P < 0.01, M.S.E. = 8531.24. P3 latency for respond trials (364 msec) was significantly shorter
TABLE Mean
than P3 latency for withhold trials (385 msec), F (1, 42) = 11.17, P < 0.01, M.S.E. = 2701.14. However, there were no significant age x trial type interactions.
P3 amplitude As shown in Fig. 4 and Table III, there was a significant age × electrode location × trial type interaction in P3 amplitude, F (4, 8 4 ) = 3.83, P < 0.01, M.S.E. = 3.19. During respond trials, mean P3 amplitude decreased significantly with increasing age at Cz and at Pz (see Table III). However, at Fz elderly subjects as well as young subjects had significantly larger mean P3 amplitude than middle-aged subjects. Mean P3 amplitudes at Fz for young and
RESPOND
.....
II
WITHHOLD
'.-.'
~
s u s t a i n e d p o t e n t i a l a m p l i t u d e s (p.V) for performancex
e l e c t r o d e l o c a t i o n x t r i a l type.
Fast subjects
j:d ~
.~il
Slow subjects
Respond trials Fz
- 6.9
- 6.5
....
Cz
- 11.7
- 8.7
....... E l d e r l y
Pz
- 7.4
- 6.2
o"
Withhold trials Fz
- 3.1
- 2.8
Cz
- 3.4
- 3.6
Pz
- 3.5
- 3.3
'40"
;o'oo
Middle-aged
~7
_ !5pV
o'"
"soo mseg
"1ooo
F i g . 3. Grand mean e v e n t - r e l a t e d p o t e n t i a l s d u r i n g t h e non-signaled task. T h e onset of t h e l e t t e r ( i m p e r a t i v e stimulus) was a t 0 msec.
A G I N G , R E S P O N D - W I T H H O L D TASKS, A N D ERPs il 21.o~
RESPOND TRIALS
19.0~
"-..
g ,,.o1.,-
"~,,
.......j
. "-,
xs.o
L II
_.13.0 < 1
subjects had significantly larger mean P3 amplitudes than middle-aged subjects at Cz and Pz. At Fz elderly subjects produced larger mean P3 amplitude than either of the younger groups, while young subjects had larger mean P3 amplitudes than middle-aged subjects.
WITHHOLD TRIALS
".%<
o.
a
~
135
" %., %
%
,."
.
; .01.
....,,..o
Fz o,e.o Cz
% YOUNG MIDDLE ELDERLY -AGED
ez
,~..-o
•
1
1
YOUNG MIDDLE ELDERLY -AGED
Fig. 4. Mean P3 amplitudes of young, middle-aged, and elderly subjects during respond and withhold trials at Fz, Cz, and Pz.
elderly subjects did not differ significantly during respond trials. During withhold trials, young subjects had larger mean P3 amplitudes than middle-aged or elderly subjects at Cz and at Pz. However, elderly
Late activity Mean late activity, which followed the period for measurement of the P3, was positive and significantly larger during withhold trials (4.4 ffV) than during respond trials (0.3 ffV) over all subjects and electrode locations, F (1, 42)= 25.30, P < 0.001, M.S.E. = 47.40. The age x trial type interaction was significant, F (2, 42) = 7.06, P < 0.01, M.S.E. = 47.40, indicating that elderly subjects had a larger difference in late activity between respond and withhold trials than the younger groups (elderly, - 1 . 4 and 6.9 ffV; middle-aged, 2.0 and
T A B L E Ill Mean P3 amplitude (ffV) for age x electrode location x trial type and significant mean comparisons among young (Y). middle-aged (M), and elderly (E) groups. Statistics which indicate significance of the mean comparisons are shown below. Young
Middle-aged
Elderly
Mean comparisons
16.9 17.9 20.4
14.8 14.7 18.6
17.4 11.3 15.0
Y, E > M * * Y > M,E**; M > E** Y > M,E**; M > E**
16.1 18.6 19.0
14.4 13.5 14.5
17.9 15.4 15.9
E > Y,M**; Y > M * * Y > E,M**; E > M * * Y > E,M**; E > M*
Respond trials Fz Cz Pz
Withhold trials Fz Cz Pz
Young vs. middle-aged
r,~
Young vs. elderly
q
Middle-aged vs. elderly
r,~
q
r,~
q
Respond trials" Fz Cz Pz
2, 84 2, 84 2, 84
4.63 7.12 4.20
3, 84 3, 84
14.67 12.15
3, 84 2, 84 2, 84
5.83 7.55 7.95
2, 84 3, 84 3, 84
3.83 11.32 9.96
2, 84 2, 84 2, 84
4.13 7.17 6.75
3, 84 2, 84 2, 84
7.96 4.15 3.21
Withhold trials Fz Cz Pz * P < 0.05; ** P < 0.01.
136
J.A. P O D L E S N Y Fz
cz
vz
~. - 4 . 0
r. F--
•
".
0.0
~+4o
11
o..
÷s0
;~'o RESPOND
WITHHOLD
RESPOND TRIAL
WITHHOLD TYPE
RESPOND
I
WITHHOLD
Fig. 5. Mean late activity of young, middle-aged, and elderly subjects during respond and withhold trials at Fz, Cz, a n d Pz.
3.6/~V; young, 0.4 and 2.7/~V, respectively). The age × electrode location × trial type interaction was also significant, F (4, 84)= 5.84, P < 0.001, M.S.E. = 4.81. As shown in Fig. 5 and Table IV, the pattern of mean late activity for the different age groups and trial types varied across electrode locations. During respond trials at Fz,
ET AL.
young subjects had a larger negative mean than the middle-aged and elderly subjects, while the middle-aged and elderly subjects did not differ significantly. During withhold trials at Fz, mean late activity showed significantly larger positive values with increasing age (see Fig. 5 and Table IV). At Cz, elderly subjects produced the largest negative mean late activity during respond trials and the largest mean positive late activity during withhold trials. The Cz means for the young and middle-aged groups differed significantly for respond trials but not for withhold trials. All of the means were positive at Pz. The means for the young and middle-aged groups did not differ significantly for either respond or withhold trials. However, at Pz the elderly group had significantly smaller positive mean late activity during respond trials and significantly larger positive mean late activity during withhold trials.
T A B L E IV
Mean late activity (p,V) for age × electrode location x trial type and significant mean comparisons among young (Y), m i d d l e - a g e d (M), and elderly (E) groups. Statistics which indicate significance of the mean comparisons are shown below. Young
Middle-aged
Elderly
Mean comparisons
- 3.3 -1.9 6.4
- 1.3 0.3 7.0
- 1.2 -7.5 4.4
E,M > Y * Y,M > E**: M > Y** M > E * * : Y > E*
0.8 2.0 5.4
2.9 3.0 4.9
7.4 5.9 7.5
E > Y,M**: M > Y** E > Y,M** E > Y,M**
Respondtrials Fz Cz Pz
WithhoM trials" Fz Cz Pz
Young vs. middle-aged
Young vs. elderly
Middle-aged vs. elderly
r,~
q
r,~
q
r,~
q
2, 84 2, 84
3.66 3.95 -
3, 84 2, 84 2, 84
3.68 10.22 3.65
3, 84 3, 84
14.18 4.85
2, 84
3.97 -
3, 84 3, 84 2, 84
12.05 7.18 3.84
2, 84 2, 84 3, 84
8.09 5.29 4.79
Respondtrials Fz Cz Pz
Withhold trials Fz Cz Pz * P < 0.05; ** P < 0.01.
-
AGING, RESPOND-WITHHOLD TASKS, AND ERPs
f
ELDERLY SUBJECTS, Cz
7
Respond trials ....... Withhold trial,
.: / •
0
T.
-42 -37 -32 -27 -22 -17 -12 -7 -2 3 8 113 18 MEAN LATE ACTIVITY (INTERVAL MIDPOINTS,pV)
Fig. 6. Frequency distributions comparing the mean amplitude of late activity for elderly subjects during respond and withhold trials at ('z.
To determine if the unexpectedly large negative respond and positive withhold LA means of elderly subjects at Cz were due to a few extreme values, the moment coefficients of skewness were computed for these distributions. Both distributions were moderately negatively skewed (respond trials, y~ = -0.83; withhold trials, Yt = -1.33). As may be seen in Fig. 6, the distribution for respond trials contained predominantly negative values, and the distribution for withhold trials contained predominantly positive values. Correlations between SP and LA
Since the large negative SP and LA values at Cz during respond trials in the elderly group suggested a common process, we computed correlation coefficients between SP amplitude and LA for respond trials at Cz for the different age groups. SP amplitude was significantly related to LA in the elderly group ( r = 0.51, P < 0.05) but not in the young and middle-aged groups (rs = 0.08 and 0.31, respectively).
Discussion
The hypothesis that inhibitory activity might produce more positive potentials during withhold trials was generally supported by the results with LA but not by those with P3 amplitude. LA at Fz and Cz shifted from values which were generally negative during respond trials to values which
137
were positive during withhold trials. Surprisingly, this effect was strongest with elderly subjects. However, there was no definite trend toward larger P3 amplitude during withhold trials. This finding with P3 amplitude is contrary to that of Karlin et al. (1970). The results with SP amplitude are equivocal with respect to this hypothesis, since the means were negative for withhold trials, as well as respond trials. The fact that withhold trials were significantly less negative than respond trials might be interpreted to indicate a summative positive component or simply less negative activity for trials during which the subjects were not required to respond. Certain of the results may be interpreted as indicating weakened inhibitory activity in the elderly. The elderly group produced the largest negative SPs and LA at Cz during respond trials and this might be explained by stronger excitation due to weakened inhibition. However, the elderly group also produced the most positive LA during withhold trials at all electrode sites. If it is assumed that positive potentials reflect inhibitory activity, the results with LA suggest stronger inhibitory activity in the elderly subjects. Alternatively, the large SPs and LA in the elderly group could indicate weakened inhibitory control of processes which give rise to these potentials. There were several puzzling results. Elderly subjects produced the largest negative SPs at Cz during respond trials. This agrees with our previous finding in a different sample of subjects (Podlesny and Dustman 1982) and appears paradoxical since in both studies, faster performing subjects produced larger negative SPs while the elderly groups performed most slowly. The large negative LA which appeared at Cz during respond trials in the elderly group was unexpected and was related to the large negative SPs in that group. The large positive LA during withhold trials in the elderly subjects is also surprising. Finally, the reduction of P3 amplitude with age at Pz and Cz parallels previous findings (Tecce et al. 1980, 1982; Podlesny and Dustman 1982). Tecce and coworkers have also noted that frontal P3 amplitude was not reduced in elderly people, and they suggested that this may indicate strength of orienting a n d / o r dishabituation processes in the elderly (Tecce et al.
138 1982). The present study produced larger frontal P3 amplitude in the elderly group, particularly during withhold trials. The shift in maximal P3 amplitude from Cz and Pz in the young group to Fz in the elderly group suggests a more frontal distribution of function in the elderly. The present results are subject to a variety of interpretations of which a few might be suggested. Aside from the question of whether inhibitory processes are reflected in these late potentials, it is noteworthy that the elderly people produced significantly larger negative SPs at Cz during respond trials, significantly larger negative and positive LA at Cz during respond and withhold trials, respectively, and significantly larger P3 amplitudes at Fz during withhold trials relative to the younger groups. These findings suggest the possibility that the elderly subjects may have applied greater effort to performance on the tasks. Perhaps adequate performance of the tasks was more challenging for the elderly and required them to mobilize neural resources to a greater extent. Such mobilization of (as yet indefinite) processes could be the source of their relatively large potentials. It may be suggested that more difficult respond-withhold tasks might also enhance these potentials in younger subjects. Since all subjects were selected for good health, these results may also be interpreted as arising from possible differential effects of the stringent selection criteria for our subjects. The elderly group may represent 'survivors,' individuals who have achieved old age in good health (Siegler et al. 1980). Large potentials among such healthy elderly subjects might reflect adaptive mechanisms which may have been innate or achieved developmentally, possibly in response to the decline of other functions.
Summary Differences in late event-related potentials (ERPs) during active respond-withhold tasks were tested with 16 young, 16 middle-aged, and 16 elderly subjects (mean ages were 30, 50 and 69 years, respectively). All subjects were selected for good health. While monopolar electroencephalograms were recorded from Fz, Cz, and Pz, subjects
J.A. PODLESNY ET AL. performed a signaled respond-withhold task to allow measurement of sustained potentials (SPs) and a non-signaled respond-withhold task for measurement of P3 responses and late activity (LA) which follows the P3. Respond and withhold trials were presented randomly with a probability of 0.5 in both tasks. Compared with younger subjects, the elderly group produced significantly larger negative mean SP amplitude during respond trials at Cz, significantly larger negative mean LA during respond trials at Cz, and significantly larger positive LA during withhold trials at all electrode locations. The elderly group also produced significantly larger P3 amplitude at Fz. The results are discussed in relation to the hypothesis that inhibition may be weaker in elderly subjects. Two further possible interpretations of the larger ERPs in elderly subjects are suggested: (1) the elderly group may have been more challenged by the tasks and their larger ERPs may reflect greater neural effort; a n d / o r (2) the health selection procedures may have produced a group of elderly 'survivors' whose large ERPs may be characteristic of robust individuals or may reflect adaptations to aging-related changes in neural function.
Rdsum6 Vieillissement et thche d'arrOt de r~ponse." effets sur les potentiels permanents, les rkponses P3 et I'activitb tardive
On a recherch6 des diff6rences entre potentiels tardifs li6s ~ l'6v6nement d6velopp6s au cours de t~che active de r6ponse ou d'abstention de r6ponse, chez 16 jeunes, 16 sujets d'~ge moyen et 16 sujets plus ~g6s (ages moyens 30, 50 et 69 ans respectivement). Tousles sujets s61ectionn6s 6taient en bonne sant6. Pendant que les sujets effectuaient une telle thche signal6e, un EEG monopolaire 6tait recueilli en Fz, Cz et Pz et l'on mesurait le potentiel stable (PS); une thche analogue, non signal6e, servait 6valuer les r6ponses P3 et l'activit6 tardive subs6quente. Les essais avec r6ponse ou non-r6ponse 6taient pr6sent6s al6atoirement, avec pour chacun une probabilit6 de 0,5.
AGING, RESPOND-WITHHOLD TASKS. AND ERPs
Par rapport h des sujets plus jeunes, les sujets ~g6s ont d6velopp6 une activit6 tardive moyenne n6gative significativement plus grande pendant les essais avec r6ponse en Cz, de m6me qu'une activit6 tardive positive plus grande pendant les essais ou la r6ponse 6tait emp6ch6e, ceci sur toutes les 61ectrodes. Le groupe le plus fi.g6 a 6galement pr6sent6 une P3 d'amplitude plus grande h Fz. Ces r6sultats sont discut6s en relation avec l'hypoth6se de la possibilit6 d'une inhibition plus faible chez les sujets ~tg6s. Deux autres interpr6tations possibles sont sugg6r6es: (1) le groupe ~g6 s'est peut 6tre plus investi dans les tests et l'ampleur de leurs potentiels pourrait refl6ter un plus grand effort c6r6bral; et/ou (2) le crit6re de s61ection par la sant6 a produit un groupe ag6 de 'survivants' pour lesquels l'ampleur des potentiels 6voqu6s pourrait 6tre caract6ristique d'individus robustes ou pourrait refl6ter des adaptations h des modifications de la fonction neuronale li6es au vieillissement. The technical assistance provided by Albert Johnson and Dr. Cameron Schlehuber is gratefully acknowledged.
References Adolfsson, R., Gottfries, C.G., Roos, B.E. and Winblad, B. Post mortem distribution of dopamine and homovanillic acid in human brain, variations related to age, and a review of the literature. J. neural Transm., 1979, 45: 81-105. Agrawal, H.C. and Himwich, W.A. Amino acids, proteins and monoamines of developing brain. In: W.A. Himwich (Ed.), Developmental Neurobiology. Thomas, Springfield, IL, 1970: 287-310. Beck, C.H.M. Functional implications of changes in the senescent brain: a review. Canad, J. neurol. Sci., 1978, 5: 417-424. Brobeck, J.R. Best and Taylor's Physiological Basis of Medical Practice, 9th edn. Williams and Wilkins, Baltimore, MD, 1973. Dustman, R.E. and Beck, E.C. The effects of maturation and aging on the waveform of visually evoked potentials. Electroenceph, clin. Neurophysiol., 1969, 26: 2-11. Dustman, R.E. and Snyder, E.W. Life-span changes in visually evoked potentials at central scalp. Neurobiol. Aging, 1981, 2: 303-308. Dustman, R.E., Snyder, E.W. and Schlehuber, C.J. Life-span alterations in visually evoked potentials and inhibitory function. Neurobiol. Aging, 1981, 2: 187-192. Frolkis, V.V. and Bezrukov, V.V. Aging of the Central Nervous System. Interdisciplinary Topics in Gerontology, Vol. 16. Karger, New York, 1979:131 pp.
139 Howard, R., Fenton, G.W. and Fenwick, P.B. Slow cerebral potentials in a 'go-no go' avoidance situation: a study on special hospital patients. In: C. Barber (Ed.), Evoked Potentials. University Park Press, Baltimore, MD, 1980: 515-523. Jakubczak, L.F. Age and animal behavior. In: C. Eisdorfer and M.P. Lawton (Eds.), The Psychology of Adult Development and Aging. American Psychological Association, Washington, DC, 1973: 98-111. Karlin, L., Martz, M.J. and Mordkoff, A.M. Motor performance and sensory-evoked potentials. Electroenceph. clin. Neurophysiol., 1970, 28: 307-313. McCarthy, G. and Donchin, E. The effect of temporal and event uncertainty in determining the waveforms of the auditory event related potential (ERP). Psychophysiology, 1976, 13: 581-590. McGeer, P.L. and McGeer, E.G. Chemistry of mood and emotion. Ann. Rev. Psychol., 1980, 31: 273-307. Podlesny, J.A. and Dustman, R.E. Age effects on heart rate, sustained potential, and P3 responses during reaction-time tasks. Neurobiol. Aging, 1982, 3: 1-9. Pradhan, S.N. Minireview: central neurotransmitters and aging. Life Sci., 1980, 26: 1643-1656. Rohrbaugh, J.W., Syndulko, K. and Lindsley, D.B. Brain wave components of the contingent negative variation in humans. Science, 1976, 191: 1055-1057. Rohrbaugh, J.W., Syndulko, K., Sandquist, T.F. and Lindsley, D.B. Synthesis of the contingent negative variation brain potential from noncontingent stimulus and motor elements. Science, 1980, 208: 1165-1168. Sandquist, T.F., Rohrbaugh, J.W., Syndulko, K. and Lindsley, D.B. Electrocortical signs of levels of processing: perceptual analysis and recognition memory. Psychophysiology, 1980, 17: 568-576. Scheibel, M.G. and Scheibel, A.B. Structural changes in the aging brain. In: H. Brody, D. Harmon and J.M. Ordy (Eds.), Aging, Vol. 1. Raven Press, New York, 1975: 11-37. Siegler, I.C., Nowlin, J.B. and Blumenthal, J.A. Health and behavior: methodological considerations for adult development and aging. In: L.W. Poon (Ed.), Aging in the 1980s. American Psychological Association, Washington, DC, 1980: 599-612. Simson, R., Vaughn, Jr., H.G. and Ritter, W. The scalp topography of potentials in auditory and visual g o / n o go tasks. Electroenceph. clin. Neurophysiol., 1977, 43: 864-875. Straumanis, J.J., Shagass, C. and Overton, D.A. Somatosensory evoked responses in Down's syndrome. Arch. gen. Psychiat., 1973, 29: 544-549. Tecce, J.J., Yrchik, D.A., Meinbresse, D., Dessonville, C.L. and Cole, J.O. CNV rebound and aging. I. Attention functions. In: H.H. Kornhuber and L. Deecke (Eds.), Progress in Brain Research. Vol. 54. Motivation, Motor and Sensory Processes of the Brain: Electrical Potentials, Behavior and Clinical Use. Elsevier/North-Holland Biomedical Press, Amsterdam, 1980: 552-561. Tecce, J.J., Cattanach, L., Yrchik, D.A., Meinbresse, D. and Dessonville, C.L. CNV rebound and aging. Electroenceph. clin. Neurophysiol., 1982, 54: 175-186.