Brain Research, 521 (1990) 281-288
281
Elsevier BRES 15675
Gating of somatosensory input by human prefrontal cortex S. Yamaguchi 1'2 and R.T. Knight 1 ~Department of Neurology, University of California, Davis, Veterans Administration Medical Center, Martinez, CA 94553 (U.S.A.)and 2Third Division of Internal Medicine, Shimane Medical University, Izumo (Japan) (Accepted 9 January 1990)
Key words: Prefrontal cortex; Somatosensory evoked potential; Gating; Inhibition
Somatosensory evoked potentials (SEPs) to median nerve stimulation were recorded in controls and in patients with focal lesions in dorsolateral prefrontal cortex (PFCx). Unilateral PFCx lesions increased the amplitude of the P26 component generated in postcentral areas 1 and 2. The amplitudes of the N28, P45 and N67 SEP components recorded over post-rolandic and frontal electrodes were also enhanced by PFCx damage. In contrast, the N19 component generated in postcentral area 3b was unaffected by PFCx lesions. The results indicate that PFCx exerts inhibitory modulation on sensory processing that may be mediated by corticocortical PFCx-parietal connections. INTRODUCTION
MATERIALS AND METHODS
Bilateral prefrontal cortex (PFCx) damage results in distinctive behavioral deficits including distractibility, poor attention capacity, perseveration and imitation behavior 3'9A°'17'21'3°. Lhermitte and coworkers attributed
Subjects and patients
some of these disturbances to disinhibition of parietal sensorimotor control mechanisms secondary to frontal damage27,28 Electrophysiological studies reveal a net inhibitory PFCx output to cortical and subcortical structures 1,16,38. In animals, somatosensory input is gated at several levels extending from the periphery 26 through dorsal column nuclei6'18 and thalamus 41 to the sensory cortex 4'47. In cats, a PFCx-nucleus reticularis thalami system modulates input to somatosensory cortex, with reversible cryogenic suppression of PFCx resulting in e n h a n c e m e n t of primary cortical activity38. Unilateral PFCx damage in h u m a n also impairs gating of auditory input to primary cortical regions 25. Patients with cortical reflex myoclonus have enhanced somatosensory evoked potential (SEP) amplitudes 15,23. A m p l i t u d e increases are observed for the parietal P25 and frontal N30 c o m p o n e n t s but not for the parietal N20 and frontal P22 SEPs. These results suggest that neural elements in somatosensory cortex may be differentially affected by excitatory or inhibitory inputs. In the current study, we examined the role of human PFCx in modulation of different regions of somatosensory cortex by recording SEPs in patients with unilateral PFCx damage.
Two age matched groups of 8 controls and 8 patients were studied. All gave informed consent and were paid for participating. The control group consisted of neurologically normal right-handed subjects (7 male, 1 female; mean age 62 + 13 years) recruited from hospital staff personnel. The patient group (right handed; 7 male, 1 female; 7L, 1R; mean age 66 + 7 years) was selected on the basis of having a unilateral prefrontal lesion on CT scan and no evidence for sensory deficits by history or clinical examinations. All patients were studied at least 6 months postlesion and included CVA (n = 7) and encephalomalacia status-post meningioma resection (n = 1). A few patients had mild non-fluent aphasia (n = 2) or hand weakness (n = 1). All patients were living independently and had neither the classic frontal syndrome seen in bilateral PFCx damage nor psychological problems (e.g. depression or anxiety). Lesions were reconstructed from CT scans and centered in Brodman area 9 and 46, with some patients showing extension into area 4, 6, 8, 44 and 45 (see Fig. 1).
Recording techniques SEPs were recorded in a dimly lit, electrically and acoustically shielded room. The patient was seated confortably in a reclining armchair and encouraged to relax. The stimuli were square-wave pulses of 0.15 ms duration delivered to the median nerve at the wrist through Ag/AgCI electrodes. Stimulus intensity was set at 10% above twitch threshold at a rate of 3 Hz. Ag/AgCI electrodes were placed at F3, F4, C3p , C4p , P3, P4, T3, T 4, T 5, T 6, Fz, Czp and Pz on the scalp (International 10-20 system). C3p, C4p and Czp were located 2 cm posterior to C3, C4 and Cz, respectively. For convention, the C3p or C4p and P3 o r P4 electrodes contralateral to the hand of stimulation will be referred to as C¢ and Pc (c = contralateral). Additional electrodes were placed at Erb's point above the clavicle overlying the brachial plexus (Bp) and at the posterior midline of the neck at the second cervical vertebra (Cv2). Linked earlobe electrodes were used as the reference and a ground electrode was attached to the forehead. Electrode impedances were kept below 5 kg2 and bandpass was set at 10-1000 Hz.
Correspondence: S. Yamaguchi, Department of Neurology, University of California, Davis, Veterans AdministrationMedical Center, 150 Muir Road, Martinez, CA 94553, U.S.A. 0006-8993/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
282
6°
'
1.
3
8
:
4
.
5
' "e
"-
7
t O 0 ; ~.
33~
f
Fig. 1. Lesion reconstruction in patients with unilateral damage in prefrontal cortex (L = 7, R = 1; mean volume = 38.8 cm3). Lesions are all reflected onto the left side of the figures. The lines on the lateral reconstruction indicate the location of the corresponding axial section. The scale indicates the percentage of patients with damage in the corresponding area.
A 331-ms epoch was digitized on-line to magnetic tape at a sampling rate of 769 points/s. Trials with excess E M G interference were automatically excluded from the average. A total of 600-800 responses were averaged with replications of an additional 600-800 responses for each hand of stimulation in a counterbalanced design. Each data block required about 5 min of stimulation. Subject arousal levels were monitored by observation through closed-circuit video. Subjects were quietly seated with their head erect and their eyes open during SEP recording. The experimenter talked with the subjects in the 1-2 rain inter-block rest periods. Peak SEP amplitudes were measured relative to the mean baseline voltage during the 50 ms of prestimulus activity. Peak latency was tabulated relative to stimulus delivery in the windows centered around the group mean latencies. P13, N19, P26, P40 and N67 components were identified at post-rolandic electrodes and P13, P20, N28 and P45 components were recorded at frontal sites. Measurement windows were 17-22 ms for the N19, 22-30 ms for the P26, 36-50 ms for the P40, 60-80 ms for the N67 at post-rolandic
sites and 18-23 ms for the P20, 20-40 ms for the N28, and 40-50 ms for the P45 at frontal sites. Windows were determined by assessment of individual subject and group averages.
Statistical analysis Six post-rolandic electrodes (Czp, Pz and 4 electrodes over the post-rolandic scalp contralateral to the hand of stimulation) and all 3 frontal electrodes were statistically analyzed. The cortical SEP components on the post-rolandic scalp ipsilateral to the hand of stimulation were not well seen and were not included in the analysis. Data were subjected to repeated measures analysis of variance (subject × group x side of stimulation x electrode). Since multiple electrodes do not provide independent amplitude or latency measures of components, measurements at single scalp sites (i.e. Pc, F c or Fz) were also used to evaluate reliable amplitude and latency changes in components. Isopotential maps were constructed from the normalized voltages for each subject. An automated interpolation algorithm calculated
283
a •
bll
CONTRALATERAL
IPSILATERAL N28
N28
P20 P45
P20 P45
He7
r
...
~:, ~,,-=.-.--~-N67
N19
T3
/"~ T4 ~
~
* ~
C3p~,,.ll L'!.1~../. . . . . . . .-..-,
'
,
,.,,~,-
,
l;q:!
.
~
/~ ...... .....
P26
P3
-
'
//""
T6
~V 160+
: :0: : : : ' ' : : : : :80 ':"
~-~/~'~'~-~.'.-'~'~
CONTROL
FRotcrRL ......................................
Fig. 2. Grand average SEPs from the control and frontal subjects. For the patients the data are shown for stimuli to the hand contralateral (a) or ipsilateral (b) to the lesion. The control data are shown for right- (a) and left- (b) hand stimulation. Note the e n h a n c e m e n t of the P26 and N67 at post-rolandic electrodes and the N28 and P45 at frontal electrodes.
the contribution of each electrode on each interelectrode point on the map, with the contribution from each inversely weighed as a function of the cube of the interelectrode distance. The maps for the N19, P26, P40 and N67 components generated by stimulation of the hand contralateral to the lesion were compared to the right hand of controls.
Wisconsin Card Sorting Test The Wisconsin Card Sorting Test (WCST) was performed using standard methods described in the literature 31.
RESULTS TABLE I
Peak amplitude o f SEP components Values are m e a n ± S.D. ~ V ) over all post-rolandic or frontal scalp sites.
Electrode site
Control group
Frontal group
Bp Cv 2 P13
-1.3±1.5 -1.4±0.8 1.5±1.1
-1.1±1.4 -1.4±0.9 1.4±1.0
Post-rolandic
N19 P26 P40 N67
-1.8±1.1 1.2±1.2 2.8±2.3 -2.2±2.3
-1.2±1.7 3.7±5.0 3.2±2.6 -4.1±2.5
Frontal
P20 N28 P45
0.4 + 1.0 0.9 + 1.0 - 1 . 8 + 1.2 - 3 . 3 + 2.0 1.7 + 1.1 2.6 + 1.2
F-value
0.08 0.01 0.002
P
n.s. n.s. n.s.
2.56 5.89 0.54 11.90
n.s.
281
Elsevier BRES 15675
Gating of somatosensory input by human prefrontal cortex S. Yamaguchi 1'2 and R.T. Knight 1 ~Department of Neurology, University of California, Davis, Veterans Administration Medical Center, Martinez, CA 94553 (U.S.A.)and 2Third Division of Internal Medicine, Shimane Medical University, Izumo (Japan) (Accepted 9 January 1990)
Key words: Prefrontal cortex; Somatosensory evoked potential; Gating; Inhibition
Somatosensory evoked potentials (SEPs) to median nerve stimulation were recorded in controls and in patients with focal lesions in dorsolateral prefrontal cortex (PFCx). Unilateral PFCx lesions increased the amplitude of the P26 component generated in postcentral areas 1 and 2. The amplitudes of the N28, P45 and N67 SEP components recorded over post-rolandic and frontal electrodes were also enhanced by PFCx damage. In contrast, the N19 component generated in postcentral area 3b was unaffected by PFCx lesions. The results indicate that PFCx exerts inhibitory modulation on sensory processing that may be mediated by corticocortical PFCx-parietal connections. INTRODUCTION
MATERIALS AND METHODS
Bilateral prefrontal cortex (PFCx) damage results in distinctive behavioral deficits including distractibility, poor attention capacity, perseveration and imitation behavior 3'9A°'17'21'3°. Lhermitte and coworkers attributed
Subjects and patients
some of these disturbances to disinhibition of parietal sensorimotor control mechanisms secondary to frontal damage27,28 Electrophysiological studies reveal a net inhibitory PFCx output to cortical and subcortical structures 1,16,38. In animals, somatosensory input is gated at several levels extending from the periphery 26 through dorsal column nuclei6'18 and thalamus 41 to the sensory cortex 4'47. In cats, a PFCx-nucleus reticularis thalami system modulates input to somatosensory cortex, with reversible cryogenic suppression of PFCx resulting in e n h a n c e m e n t of primary cortical activity38. Unilateral PFCx damage in h u m a n also impairs gating of auditory input to primary cortical regions 25. Patients with cortical reflex myoclonus have enhanced somatosensory evoked potential (SEP) amplitudes 15,23. A m p l i t u d e increases are observed for the parietal P25 and frontal N30 c o m p o n e n t s but not for the parietal N20 and frontal P22 SEPs. These results suggest that neural elements in somatosensory cortex may be differentially affected by excitatory or inhibitory inputs. In the current study, we examined the role of human PFCx in modulation of different regions of somatosensory cortex by recording SEPs in patients with unilateral PFCx damage.
Two age matched groups of 8 controls and 8 patients were studied. All gave informed consent and were paid for participating. The control group consisted of neurologically normal right-handed subjects (7 male, 1 female; mean age 62 + 13 years) recruited from hospital staff personnel. The patient group (right handed; 7 male, 1 female; 7L, 1R; mean age 66 + 7 years) was selected on the basis of having a unilateral prefrontal lesion on CT scan and no evidence for sensory deficits by history or clinical examinations. All patients were studied at least 6 months postlesion and included CVA (n = 7) and encephalomalacia status-post meningioma resection (n = 1). A few patients had mild non-fluent aphasia (n = 2) or hand weakness (n = 1). All patients were living independently and had neither the classic frontal syndrome seen in bilateral PFCx damage nor psychological problems (e.g. depression or anxiety). Lesions were reconstructed from CT scans and centered in Brodman area 9 and 46, with some patients showing extension into area 4, 6, 8, 44 and 45 (see Fig. 1).
Recording techniques SEPs were recorded in a dimly lit, electrically and acoustically shielded room. The patient was seated confortably in a reclining armchair and encouraged to relax. The stimuli were square-wave pulses of 0.15 ms duration delivered to the median nerve at the wrist through Ag/AgCI electrodes. Stimulus intensity was set at 10% above twitch threshold at a rate of 3 Hz. Ag/AgCI electrodes were placed at F3, F4, C3p , C4p , P3, P4, T3, T 4, T 5, T 6, Fz, Czp and Pz on the scalp (International 10-20 system). C3p, C4p and Czp were located 2 cm posterior to C3, C4 and Cz, respectively. For convention, the C3p or C4p and P3 o r P4 electrodes contralateral to the hand of stimulation will be referred to as C¢ and Pc (c = contralateral). Additional electrodes were placed at Erb's point above the clavicle overlying the brachial plexus (Bp) and at the posterior midline of the neck at the second cervical vertebra (Cv2). Linked earlobe electrodes were used as the reference and a ground electrode was attached to the forehead. Electrode impedances were kept below 5 kg2 and bandpass was set at 10-1000 Hz.
Correspondence: S. Yamaguchi, Department of Neurology, University of California, Davis, Veterans AdministrationMedical Center, 150 Muir Road, Martinez, CA 94553, U.S.A. 0006-8993/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
282
6°
'
1.
3
8
:
4
.
5
' "e
"-
7
t O 0 ; ~.
33~
f
Fig. 1. Lesion reconstruction in patients with unilateral damage in prefrontal cortex (L = 7, R = 1; mean volume = 38.8 cm3). Lesions are all reflected onto the left side of the figures. The lines on the lateral reconstruction indicate the location of the corresponding axial section. The scale indicates the percentage of patients with damage in the corresponding area.
A 331-ms epoch was digitized on-line to magnetic tape at a sampling rate of 769 points/s. Trials with excess E M G interference were automatically excluded from the average. A total of 600-800 responses were averaged with replications of an additional 600-800 responses for each hand of stimulation in a counterbalanced design. Each data block required about 5 min of stimulation. Subject arousal levels were monitored by observation through closed-circuit video. Subjects were quietly seated with their head erect and their eyes open during SEP recording. The experimenter talked with the subjects in the 1-2 rain inter-block rest periods. Peak SEP amplitudes were measured relative to the mean baseline voltage during the 50 ms of prestimulus activity. Peak latency was tabulated relative to stimulus delivery in the windows centered around the group mean latencies. P13, N19, P26, P40 and N67 components were identified at post-rolandic electrodes and P13, P20, N28 and P45 components were recorded at frontal sites. Measurement windows were 17-22 ms for the N19, 22-30 ms for the P26, 36-50 ms for the P40, 60-80 ms for the N67 at post-rolandic
sites and 18-23 ms for the P20, 20-40 ms for the N28, and 40-50 ms for the P45 at frontal sites. Windows were determined by assessment of individual subject and group averages.
Statistical analysis Six post-rolandic electrodes (Czp, Pz and 4 electrodes over the post-rolandic scalp contralateral to the hand of stimulation) and all 3 frontal electrodes were statistically analyzed. The cortical SEP components on the post-rolandic scalp ipsilateral to the hand of stimulation were not well seen and were not included in the analysis. Data were subjected to repeated measures analysis of variance (subject × group x side of stimulation x electrode). Since multiple electrodes do not provide independent amplitude or latency measures of components, measurements at single scalp sites (i.e. Pc, F c or Fz) were also used to evaluate reliable amplitude and latency changes in components. Isopotential maps were constructed from the normalized voltages for each subject. An automated interpolation algorithm calculated
283
a •
bll
CONTRALATERAL
IPSILATERAL N28
N28
P20 P45
P20 P45
He7
r
...
~:, ~,,-=.-.--~-N67
N19
T3
/"~ T4 ~
~
* ~
C3p~,,.ll L'!.1~../. . . . . . . .-..-,
'
,
,.,,~,-
,
l;q:!
.
~
/~ ...... .....
P26
P3
-
'
//""
T6
~V 160+
: :0: : : : ' ' : : : : :80 ':"
~-~/~'~'~-~.'.-'~'~
CONTROL
FRotcrRL ......................................
Fig. 2. Grand average SEPs from the control and frontal subjects. For the patients the data are shown for stimuli to the hand contralateral (a) or ipsilateral (b) to the lesion. The control data are shown for right- (a) and left- (b) hand stimulation. Note the e n h a n c e m e n t of the P26 and N67 at post-rolandic electrodes and the N28 and P45 at frontal electrodes.
the contribution of each electrode on each interelectrode point on the map, with the contribution from each inversely weighed as a function of the cube of the interelectrode distance. The maps for the N19, P26, P40 and N67 components generated by stimulation of the hand contralateral to the lesion were compared to the right hand of controls.
Wisconsin Card Sorting Test The Wisconsin Card Sorting Test (WCST) was performed using standard methods described in the literature 31.
RESULTS TABLE I
Peak amplitude o f SEP components Values are m e a n ± S.D. ~ V ) over all post-rolandic or frontal scalp sites.
Electrode site
Control group
Frontal group
Bp Cv 2 P13
-1.3±1.5 -1.4±0.8 1.5±1.1
-1.1±1.4 -1.4±0.9 1.4±1.0
Post-rolandic
N19 P26 P40 N67
-1.8±1.1 1.2±1.2 2.8±2.3 -2.2±2.3
-1.2±1.7 3.7±5.0 3.2±2.6 -4.1±2.5
Frontal
P20 N28 P45
0.4 + 1.0 0.9 + 1.0 - 1 . 8 + 1.2 - 3 . 3 + 2.0 1.7 + 1.1 2.6 + 1.2
F-value
0.08 0.01 0.002
P
n.s. n.s. n.s.
2.56 5.89 0.54 11.90
n.s.
