Varying expressions of alerting mechanisms in wakefulness and across sleep states

Electroencephalography and clinical Neurophysiology, 82 (1992) 458-468 © 1992 Elsevier Scientific Publishers Ireland, Ltd. 0013-4649/92/$05.00

458

EEG 91095

Varying expressions of alerting mechanisms in wakefulness and across sleep states L a r r y D. S a n f o r d a, A d r i a n R. M o r r i s o n a,b,c W i l l i a m A. Ball b, R i c h a r d J. R o s s b,c a n d G r a z i e l l a L. M a n n a '~Laboratories of Anatomy, School of Veterinary Medicine, b Department of PbTchiatry, School of Medicine, and c Institute of Neurological Sciences, Unicersity of Pennsyh,ania, Philadelphia, PA 19104-6045 (U.S.A.) (Accepted for publication: 14 January 1992)

Summary

Alerting stimuli, such as intense tones, presented to cats in wakefulness (W) elicit the orienting response (OR) a n d / o r the acoustic startle reflex (ASR) in conjunction with elicited ponto-geniculo-occipital waves (PGO E) from the lateral geniculate body (LGB) and elicited waves from the thalamic central lateral nucleus (CLE). Alerting stimuli presented during rapid eye movement sleep (REM) and non-rapid eye movement sleep (NREM) also elicit PGO E. We presented tones in W, REM and NREM to determine whether CL E could be obtained in sleep and to examine the patterns of responsiveness of PGO E and CL E across behavioral states. Also, we recorded ASR and OR and compared the response patterns of behavioral and central correlates of alerting. The subjects were 7 cats; all exhibited spontaneously occurring waves in LGB and CL. All cats exhibited PGO E and 5 cats exhibited CL E in W, REM and NREM. PGO E and CL E showed less evidence of habituation than did ASR and OR. The pattern of responsiveness of CL E across behavioral states was different from that found for PGO E, and spontaneous CL waves were much rarer than the LGB waves. ASR was elicited in 5 cats during W trials, and in 3 cats during REM trials. OR habituated rapidly in W and did not occur in REM and NREM. The data indicate that central mechanisms of alerting function in sleep states as well as in W and suggest that CL E and PGO E reflect activity in mechanisms underlying cortical desynchronization and visual processes which may act in concert during alerting. Key words: Acoustic startle reflex; Central lateral thalamic nucleus; Elicited PGO waves; Lateral geniculate body; Orienting; Sleep

Ponto-geniculo-occipital waves (PGO) occur spontaneously, and with high frequency, in the pons, lateral geniculate body (LGB) and occipital cortex during rapid eye movement sleep (REM) and during the transition from non-rapid eye movement sleep (NREM) into REM (Jouvet 1967). In waking (W), the LGB yields waves closely associated with eye movements (EMP) (Jeannerod and Sakai 1970). In addition, PGO-like waves have been recorded in other thalamic nuclei (e.g., Hobson 1964; Hu et al. 1989b), in the amygdala (Calvo and Fernandez-Guardiola 1984), in the cerebellum (Jeannerod 1965; Pellet et al. 1974; Marks et al. 1980) and over wide areas of the cortex (Brooks 1968) suggesting widespread phasic activation in the central nervous system (CNS).

Correspondence to: Dr. Larry D. Sanford, Laboratories of Anatomy, School of Veterinary Medicine, Room M103, 3800 Spruce Street, Philadelphia, PA 19104-6045 (U.S.A.). Tel.: (215) 898-4569 or 898-7905. Supported by USPHS Grants MH42903 (A.R.M.), MH18825 (L.D.S.) and the Department of Veterans Affairs (R.J.R.).

Because tones elicit PGO-like waves (PGO E) in W, NREM and REM, Bowker and Morrison (1976, 1977) proposed that PGO are not merely passive neurophysiological concomitants of REM, but rather are markers of CNS responses to alerting stimuli generated by neurons that in W would be involved in the production of the startle reflex and orienting. This seemingly paradoxical hypothesis suggests that a portion of the CNS active in REM could be engaged in something akin to continual orienting while the startle reflex and overt behavioral orienting (OR) are somehow inhibited (Morrison 1979). Alerting stimuli, such as intense tones, presented to cats often elicit OR (Sprague et al. 1963; Ball et al. 1991b; Sanford et al. 1992) a n d / o r the acoustic startle reflex (ASR) (e.g., Wu et al. 1989; Sanford et al. 1992) in addition to electrophysiological responses in the CNS. PGO E often accompany ASR a n d / o r OR (e.g., Bowker and Morrison 1976; Wu et al. 1989; Ball et al. 1991b). Alerting stimuli presented during REM and during NREM elicit PGOE, whereas the behavioral components of alerting are greatly suppressed or absent. ASR and OR have been observed in response to tones during REM in cats with pontine lesions that eliminate muscle atonia (REM-A) (Morrison et al.

ALERTING IN SLEEP AND WAKEFULNESS

1990). Cats with R E M - A spontaneously display both PGO and overt behaviors ranging from increased muscle jerks to full locomotion and including seemingly internally directed movements of the head and eyes (Jouvet and Delorme 1965; Henley and Morrison 1974; Sastre and Jouvet 1979; Hendricks et al. 1982). These observations in conjunction with other electrophysiological signs in R E M that are indicative of an alert brain (e.g., low voltage, fast electroencephalogram (EEG), hippocampal theta) reinforce the idea that spontaneous P G O may be in some way related to alerting mechanisms. Hobson (1964) reported spontaneous PGO-like waves in thalamic intralaminar nuclei, and we have obtained elicited wave forms from the intralaminar central lateral nucleus (CL E) with latencies that fall within the range reported for P G O E and MLR (Sanford et al. 1992). The CL is a site closely tied to processes underlying cortical desynchronization of W, and R E M (e.g., Glenn and Steriade 1982), further suggesting a role for CL E mechanisms in arousal or alerting. The similarities between spontaneous waves in LGB and CL, and between P G O E and CL E in W raise the possibility that waves in the two sites are CNS markers of related neural mechanisms of alerting. We have found that the amplitude of PGO E varies with state and that there is a general tendency for greater amplitudes in R E M than in NREM. The amplitude of PGO E has generally been found to be greater in either sleep state than in W (e.g., Ball et al. 1989b, 1991a,b). Working within the framework of Bowker and Morrison's (1977) hypothesis, these central (PGO E and CL E) and peripheral (ASR and OR) responses to alerting stimuli provide, in our view, important insights into the nature of REM, which we believe can be viewed as a paradoxical state of "hyper-alerting." However, recent evidence suggests that P G O E and ASR have largely independent mechanisms as indicated by different rates of habituation and different response patterns across stimulus intensities (Wu et al. 1989). In addition, O R habituates rapidly in W, whereas PGO E amplitude remains relatively constant across repeated tone presentations (Sokolov 1969; Ball et al. 1991b; Sanford et al. 1992). We tested the responsiveness of P G O z and CL E to tone stimuli in W, N R E M and REM. Similar patterns of responses for PGO E and CL E across and within behavioral states would suggest that the two wave forms could be markers of related underlying neural mechanisms, whereas different patterns of responses would indicate at least a partial dissociation among their underlying mechanisms. We also examined the relationship between putative CNS markers (PGO E and CL E) and peripheral indicators (ASR and OR) of alerting in W and sleep.

459

Methods and materials Seven adult, mongrel, female cats were implanted with standard electrodes normally used for sleep recording. Stainless steel screw electrodes were placed in the frontal sinuses for recording eye movements (EOG) and EEG; wire electrodes were inserted into the nuchal muscles for recording the electromyogram (EMG). Tripolar stainless steel electrodes (tip separation 1 mm) were implanted in LGB bilaterally (coordinates AP: +6.0; ML: _+10; DV: +2.7; Berman 1968) for recording P G O activity. Additional tripolar electrodes were implanted in CL bilaterally (coordinates AP: + 8.5; ML: _+4.0; DV: + 3.0; Berman 1968). Tripolar electrodes increase the probability of obtaining high amplitude P G O by increasing the possible number of electrode combinations; at least one of the electrodes must be localized within the LGB (Brooks 1967a) for registering PGO. Photomicrographs showing electrode location in CL and LGB for cat L5 are given in Fig. 1. Electrode placement within CL and LGB has been confirmed in 5 cats. One cat's brain was not processed and one is still participating in experiments. All surgical procedures were performed stereotaxically under sterile conditions using halothane or pentobarbital (42 m g / k g ) anesthesia. Subcutaneous injections of nalbuphine (2.5 m g / k g ) were used to control potential post-operative pain. Tones (100 dB SPL, 4000 Hz sine wave, 90 msec duration, 5 msec rise time, 2 sec ISI) were produced with a Harmon Kardon amplifier (model PM655 Vxl) with output to a loudspeaker (JBL model 2105H) located inside the chamber. The stimulus parameters were controlled by Coulbourn modular components. The platform for measuring ASR was based on an Endevco 2217E accelerometer and has been described in detail (Sanford et al. 1992). During testing the platform was draped with a pad and the cats had no apparent problems in sleeping. The platform was housed in a chamber (73 cm × 86 cm × 89 cm) equipped with a Plexiglas door for video monitoring. Output from the accelerometer was fed into a model 7P511J E E G amplifier in a Grass model 78D polygraph and the amplified signal was recorded on videotape along with PGOE, CL E and a stimulus marker using a Vetter model 620 recording device. For analyses, analog data were digitized using a Metrabyte Das-8 A / D converter housed in an AT compatible computer and operated by a custom program. Cats were allowed a minimum of 7 days to recover from surgery prior to being entered in any experimental protocols. Prior to testing, a series of 10 tones (1000 Hz, 90 dB) were presented to allow adjustment of the sensitivity of the PGOE, CL E and accelerometer channels for individual cats. These data points were not included in the analysis.

460

L.D. S A N F O R D ET AL.

Fig. 1. Photomicrographs of coronal sections showing electrode placement in the LGB and CL in cat L5. The sections were cut at 4 0 / z m and stained with cresyl violet. A: LGB. OT = optic tract. B: CL. MD = nucleus medialis dorsalis.

ALERTING IN SLEEP AND WAKEFULNESS

The experiment examined responses to tones in W and in sleep states. The cats were presented with 4 blocks of 40 tones in W, followed by continuous tones throughout a sleep state ( R E M or N R E M on separate days 1 week apart in a counterbalanced order) and then a final 40 tones in W. The first 4 W blocks were spaced 20 min apart and the last W block occurred 20 min after testing in the sleep state. ASR, PGO E, and CL E were recorded simultaneously and the entire experiment was videotaped. O R were scored by two independent observers from the videotaped records of cat behavior using a numerical scale adapted from Ursin et al. (1969). O R difference scores were derived from the difference between preand post-stimulus scores of head and ear movements and changes of body posture that comprise elements of behavioral O R (described in detail in Ball et al. 1991b). Elicited wave forms were defined as those with latencies between 40 and 160 msec and having a peak amplitude of at least 30% that of the mean of the 10 highest P G O E or CL z for individual cats. We obtain roughly a 90% agreement between a human observer and the computer for P G O E using this method of detection. In 2 cats the latency window for significant CL E was expanded to include the range between 20 and 160 msec due to a clear high rate of evoked potentials at approximately 30 msec post-stimulus in these animals. ASR were accepted with latencies between 40 and 200 msec. Trials with movement artifacts or ambiguous responses were dropped from the analysis. Analyses for PGOE, CL E and ASR were conducted on amplitude (mean peak amplitude of responses, excluding trials with no detectable response), magnitude (peak a m p l i t u d e / n u m b e r of stimulus presentations), latency (elapsed time to response peak after stimulus onset) and proportion (number of r e s p o n s e s / n u m b e r of stimulus presentations) in each block of 40 tones. Analyses for behavioral O R were conducted on O R difference scores (excluding trials with no discernible OR) and proportion (number of trials with O R / t o t a l number of stimulus presentations). Unless otherwise specified analyses were carried out using single factor within subjects analyses of variance (ANOVA). Post hoc tests were conducted with Tukey's HSD test with P < 0.05.

Results Overview The cats were pretested for five 6 h polygraph recording sessions to ensure that P G O spontaneously occurred in REM. All cats exhibited spontaneously occurring waves recorded in CL. The cats had experienced tonal stimuli prior to serving as subjects in the

461

present study. ASR and O R appeared to have reached a steady lower rate of responding compared to higher levels seen initially when the cats had not experienced tones. P G O E and CL E also appeared to exhibit steady rates of responding, although they showed less evidence of habituation than did ASR and O R in W. Seven cats exhibited PGOE, and 5 cats exhibited CL E in W, R E M and NREM. ASR was elicited in 5 cats during W trials, and in 3 cats during R E M trials. O R in the form of head turning and posture changes usually occurred on the first few tones in each block in W. The most prevalent O R were pinna rotations, which occurred sporadically throughout the stimulus presentations. Spontaneous and elicited activity in L G B and CL across states Elicited wave forms were obtained from the LGB and CL in all behavioral states. Obvious from the polygraph records (Fig. 2 A, B and C) is the fact that P G O E and CL z are not necessarily elicited concurrently, nor do they necessarily spontaneously co-occur in R E M and NREM. A preliminary analysis of uninterrupted 6 h sleep records indicated that spontaneous waves in CL are more frequent in N R E M than in REM; indeed, they are quite infrequent in REM. In neither sleep state do they exhibit a pattern similar to that characteristic of spontaneous P G O in R E M or in the transition from N R E M to REM, nor do they occur in bursts. The amplitude of P G O z was greater in R E M than in W, F (5, 30) = 3.487, P < 0.013, whereas the amplitude of P G O E in N R E M was not significantly different from W. In contrast, the amplitude of CL E was not significantly different from that in W in either R E M or N R E M (Fig. 3A, B). The magnitude measure for P G O E and CL E followed the same general trend as described for amplitude with the exception that the increase in P G O E magnitude in R E M relative to W did not reach significance, P < 0.097. No significant changes in proportion of trials with a response occurred in either PGOE or CLE, either within W or across behavioral state (Fig. 3C, D). No significant differences were found in latency either within W or across either sleep state for PGOE or CL z (Fig. 3E, F), though the latency of P G O z was somewhat longer in REM. Examining frequency distributions of responses for individual cats revealed that responses at particular latencies for CL E and PGO E were distributed differently within cats, and that the distribution of responses varied across states. In some cats longer latency CL z compared to PGO E occurred (Fig. 4A, D), whereas in other cats P G O E were elicited with shorter latencies than those found for CL z (Fig. 4E). The response distribution of P G O z was wider in

462

LD. SANFORD ET AL.

R E M than in N R E M (Fig. 4C, F and I), possibly due to some degree to the greater probability of a spontaneous wave falling into the extraction window. We have found that up to 20% of P G O that we accept as elicited in R E M may be spontaneous waves randomly intruding into the extraction window (Ball et al. 1989b). In all cats the post-stimulus latencies at which maximum responding of CL E and P G O E occurred were characterized by separate and distinct distributions. Kolmogorov-Smirnov 2-sample tests for cumulative frequency distributions were used to determine whether CL E and P G O E in individual cats followed the same response pattern across repeated tone presentations in the 3 test states. In W and R E M the distributions for all cats were significantly different (W: L5, P < 0.001;

L7, P < 0.001; L8, P < 0.001; L9, P < 0.001; L10, P < 0.001; L12, P < 0.005; L13, P < 0.001; REM: L5, P < 0.001; L7, P < 0.001; L9, P < 0.001; L10, P < 0.001; L13, P < 0.005). In N R E M the response patterns of CL E and P G O E were significantly different in 3 cats (L7, P < 0.001; L12 P < 0.001; L13, P < 0.002). The response patterns of CL E and P G O E in 2 cats were not significantly different (L5, P < 0.35 and L8, P < 0.08). P G O E and CL E within R E M and N R E M The trial block in R E M and N R E M generally encompassed more than 40 stimulus presentations. Changes in amplitude, magnitude, proportion and latency of CL E and P G O E were examined with 1 (sleep

A J~.._.LL~ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . EEG~r'~I--,..7

..... r ~ l ,

"r ....

r

'!

---I-""1

.... ]~-',

....

mV

J O . 6 0

i ....

," -

LGB

1.10

rnV

CL

1.10

rnV

ASR

L.

:. . . . . .

l.

;

. . . .

_lO.60rnv

_

W

B °o LGB

-1o m V

CL ASR

ST

I M

mv

_ _ . , +

"~

1.10

mV

~10.60

mv

-,

NREM

C EEG~.'~--:--:±:)::;:.7--:~-;:.::s

4.-.v~;-,.:-- ........

--~-"'v ..........

,0.60

mV

1.10

rnV

1.10

m V

,]0.60

rnv

LGB k CL ASR

8 T I M

. . . . . . . . . . . . . .

-,

~.". . . ~

.

.

.

.

:---~-~

--~

.

.

.

,----7

~

,

,-

REM Fig. 2. Polygraphrecordings from cat L5 demonstrating PGOE (in LGB) and CLE (in CL) responses to tonal stimuli. A: W. B: NREM. C: REM.

ALERTING IN SLEEP AND WAKEFULNESS

463

,. 2°°° F 1°°° I ,°°° I 500 I 0 /

C,

I

I

I

I

I

I

I

0.4 0.0 0.0

>(J z

5

I

I

I

PGO[

•

CL E I

I

0.5 0.6

E.

I

0

I

i

I

I

I

i

0.12

°I I

I

I

I

I

I

l

F.

0.09

_J°-~. o

o.o, ~

~

0.03

0.00

1

BLOCK

I

I

BLOCK

Fig. 3. Plots of (A-B) the mean amplitude (in arbitrary units), (C-D)

proportion and (E-F) latency for PGO E and CL z across states on REM and NREM test days. PGOE: N = 7. CLE: N = 5. state) × 4 (quartile) within subjects ANOVAs. Quartiles were based on the percentage time into a sleep state. The amplitude of P G O E exhibited a significant change across quartiles within R E M , F (3, 18) = 3.41, P < 0.04 and within N R E M , F (3, 15) = 3.29, P < 0.05, in each case characterized by a quadratic trend (REM: F (1, 6) = 5.92, P < 0 . 0 5 ; N R E M : F (1, 5 ) = 12.13, P < 0.018). P G O E amplitude was greater in the first quartile in R E M and N R E M , P < 0.05. When the P G O E data were considered as magnitude across quartile, the trend in R E M became linear, F (1, 6) = 9.33, P < 0.022, paralleling a downward linear trend in the proportion of responses to tonal stimuli across quartile, F (1, 6) = 12.98, P < 0.011. T h e r e were no significant differences in magnitude or proportion across quartiles in N R E M . The m e a n latencies of P G O E were significantly longer in R E M than in N R E M , F (1, 5 ) = 14.03, P < 0.013. T h e r e were no significant differences in latency across quartiles in either R E M or N R E M . The amplitude and magnitude of CL E did not significantly change across successive quartiles in REM, nor did proportion. In contrast, the A N O V A for amplitude of CL E in N R E M was significant, F (3, 1 2 ) = 8.25, P < 0.003. CL E followed a pattern similar to that found for P G O E , a significant quadratic trend, F (1, 4) = 18.99, P < 0.012, with amplitudes higher in the first and 4th quartiles, P < 0.05. No significant differ-

ences were found in magnitude, proportion or latency across quartiles in N R E M . T h e r e were no significant differences between the latencies of CL E in R E M or N R E M ; however, they did appear to be distributed differently from those of P G O E (Fig. 4). Acoustic startle and behavioral orienting across states ASR was extremely variable between cats and habituated quickly or was not observed in 4 cats. T h r e e cats exhibited consistent A S R on both the R E M and the N R E M test days. Two other cats exhibited very sporadic and infrequent A S R in W. Due to the small number of cats exhibiting A S R no statistical analysis could be conducted; however, no obvious change was found for ASR across blocks in W. The 3 cats with consistent ASR in W also startled in the R E M episode. There did appear to be a decrease in A S R amplitude across stimulus presentations in each W block of 40 tones. In REM, the amplitude of ASR in these cats was much more variable and in 2 cats did not appear to decrease within the first 40 tones. When the R E M data were considered in quartiles, 1 cat (L13) exhibited consistent A S R throughout the episode and the other (L7) exhibited A S R in the first two quartiles. A third cat (L5) startled on 3 trials in the first quartile. No significant change in O R difference scores or proportion were found across blocks of trials in W. However, within W blocks O R usually occurred on the first few trials and quickly dropped to near zero levels. As expected, no evidence of behavioral O R was found in either R E M or N R E M . We excluded from our analyses ear twitches and other phasic movements which were observed in REM. Responsiveness in W and across behavioral states Fig. 5 demonstrates the responsiveness of each variable in amplitude ( P G O E, C L z , ASR) or difference scores (OR) for 2 representative animals. These cats exhibited ASR, that occurred in conjunction with P G O E and CL E even during REM. In contrast, 4 cats exhibited little or no ASR, yet evoked wave forms were readily observed. Responsit;eness across test days To determine whether habituation occurred across test days, the data for P G O E, CL E and O R were analyzed with 2 (test d a y ) x 5 (waking block) within subjects ANOVAs. The small number of cats exhibiting A S R on both days precluded a statistical analysis on this variable; however, only 3 of 5 cats exhibiting A S R on day 1 startled on day 2. T h e r e were no obvious differences in A S R between day 1 and day 2 for the remaining cats. No significant differences between day 1 and day 2 were found for any measure of P G O E , CL E or O R and within W blocks for P G O E , CL E and

464

L.D. S A N F O R D ET AL.

W

NREM

B.

A. 4

REM C. I'~

I

nlnI1.,

0 z 0

,

D, 4 rn nbOO

PGO E

Ill c~

E.

F.

H,

I.

3

2 bJ Oq Z 0 n Oo ILl ri-

1 0

G. 4

20

160

20

150

20

160

LATENCY (MSEC) Fig. 4. Frequency histograms of P G O E and CL E during W, R E M and N R E M in selected cats demonstrating the distribution of responses at different latencies. (A-C): cat L7. (D-F): cat L8. (G-I): cat L13.

A.

l~J r~

ASR. The proportion of OR did show a significant linear decrease across W blocks on day 1, F (1, 6 ) = 11.89, P < 0.014. No significant decrease in O R across W blocks was found on day 2; however, the responsiveness across blocks on day 2 compared to day 1 was highly variable. The limited observable change in ASR and O R within and between test days was most likely due to a low stable level of responding being reached since the cats had previously served in a tone habituation study.

B.

2000

1500

p-

1000

n 500

<

I

O

I

[

i

I

i

i

0

PGO F

•

CL t i

D.

Z

C.

O.B I

9,.,.~ 0.ef 0 00 (~

0.4 0.0

Discussion

0.0

I

I

J

I

I

I

I

I

E. l °"---<)1"-'°-----°~'~ 0.12

F.

0.09

0 zb.l

0.06

t ' / -

I

ooo,

o,

~ ' ~

°;

¢

°, REM

o1 °2 °;

o',

NREM

Fig. 5. The amplitude of P G O E, CL E, A S R and O R difference scores considered as percentage of the initial block and plotted across 5 blocks of 40 tones in W on the R E M and N R E M test days for 2 selected cats. (A-B): cat L7. (C-D): cat L13.

I

The present study replicates our finding that spontaneous and elicited wave forms can be obtained from CL as well as LGB in W (Sanford et al. 1992) and demonstrates similar wave forms in R E M and NREM. The putative CNS markers of alerting mechanisms (PGO E and CL E) continue to be consistently elicited well after peripheral alerting responses (ASR and OR) have habituated within blocks in W, although sporadic O R or ASR may occur throughout a block, and OR dishabituates between blocks. This dissociation among central and peripheral mechanisms is obvious in N R E M and R E M where O R was not observed, and ASR was not observed in 4 cats.

A L E R T I N G IN SLEEP A N D W A K E F U L N E S S

P G O E were found to have significantly higher amplitudes in R E M than in W, but P G O E amplitude in N R E M was not higher than in W. P G O E in R E M occurred with noticeably longer latencies compared to NREM. These findings replicated findings in our laboratory for R E M and W, though we previously have found higher amplitude P G O E in N R E M compared to that in W (Ball et al. 1991b). This discrepancy may well reflect individual differences in a heterogeneous cat population. Our results suggest that E L E a r e not identical to PGOE, nor are spontaneous CL identical to spontaneous PGO, though they may share certain properties. CL E and PGO E were not always elicited simultaneously (the response patterns of CL E and P G O E were most similar in 2 cats in NREM), nor did spontaneous waves in LGB and CL always occur concurrently. CL E and P G O E responses were distributed at different post-stimulus latencies in the same animals in different states. In R E M and N R E M these results may have been somewhat affected by spontaneously occurring waves. Yet, we have found that only about 20% of extracted waves in R E M may be spontaneously generated (Ball et al. 1991a). Differential rates of spontaneously occurring waves did not appear to affect our results in W due to the low incidence of spontaneous waves in LGB and CL. In some cats P G O E occurred at longer latencies than CLE; in others P G O E had shorter latencies. This suggests that the two wave forms may be markers of neural mechanisms that perform different, although possibly related functions involved in alerting in the CNS. Another line of evidence that is suggestive of different functional roles for PGO E and CL E comes from their relative pattern of responses within sleep. A small, yet statistically significant, quadratic trend was found in PGO E amplitude in R E M and N R E M due to higher amplitude in the first and 4th quartiles. This finding complements a similar quadratic trend in proportion of trials with P G O E responses in R E M and N R E M reported earlier (Ball et al. 1991b). The amplitude of CL E showed a similar quadratic trend to that found for PGO E in NREM. However, CL E amplitude in R E M did not show increased amplitude compared to W, nor the change in amplitude across quartiles in R E M that was evident in PGOE. Differences between P G O E and CL E across and within states could possibly be linked to the putative functionality of the respective nuclei: visual system information gathering processes for LGB (Singer 1973, 1977; Bowker and Morrison 1977) and cortical desynchronization for CL (e.g., Morison and Dempsey 1942; Steriade 1981; Glenn and Steriade 1982). In all cases, the increases in amplitude in the 4th quartile were modest, and we did not replicate our previous finding for increased proportion of P G O E during the 4th quar-

465

tiles of R E M and N R E M in this series of cats (Ball et al. 1991b). The differences in P G O E between studies may have been due to the fact that we are dealing with subtle changes, and the analysis of sleep episodes into quartiles may simply be too gross a measure since the episodes vary so much in length. The pattern of elicited responses may vary according to whether the sleep episode ended without disruption of the sleep cycle or ended abruptly due to awakening. Also, cats differ in temperament; some are simply more sluggish than others. Further study would be needed to ascertain the variables that determine the patterns of responding of PGO E and CL E within R E M and NREM. Spontaneously occurring waves were observed in all 7 cats with electrodes in CL, whereas elicited waves were found in only 5 cats. This observation parallels previous findings and remains difficult to explain (Sanford et al. 1992). Hobson (1964) appeared to obtain dramatically different results in spontaneously occurring waves with only a slight change in electrode position within the CL. In addition, the CL is composed of two major cell groups, small neurons in the ventromedial portion and large cells in the lateral and dorsal sections (e.g., Steriade and Glenn 1982). Thus, it is possible to speculate that a distinct cell population within the CL is responsible for the activity we call CL E. Chronically implanted macroelectrodes leave large defects that make precise localization difficult (see Fig. 1). Wave-like neural activity possibly linked with spontaneous P G O is a fairly widespread phenomenon in the brain (e.g., Hobson 1964; Jeannerod 1965; Brooks 1968; Pellet et al. 1974; Marks et al. 1980; Calvo and Fernandez-Guardiola 1984; Hu et al. 1989b). Whether P G O E and CLE, and indeed other PGO-like waves, are produced by shared generator sites or reflect the operation of separate but possibly functionally related systems remains to be determined. We note that Hobson (1964) did not describe the pattern of CL activity within sleep. The present work suggests it to be quite different from that of P G O in LGB. Specifically in regard to CL E and PGOE, the pedunculo-pontine nucleus (PPN) in the brain-stem has cholinergic projections to all thalamic nuclei, including LGB and the intralaminar nuclei (Sofroniew et al. 1985; De Lima and Singer 1987; Pare et al. 1988; Smith et al. 1988; Steriade et al. 1988; Woolf et al. 1990). The peribrachial area (PB) of PPN plays a major role in P G O and EMP generation in cats; a group of PB neurons projecting to the LGB fire in bursts 15-25 msec before a P G O wave in R E M (McCarley et al. 1978; Sakai and Jouvet 1980; Nelson et al. 1983). These neurons are the last neural step in the production of PGO, if not the actual generator site (Hu et al. 1989a,b). In addition, several other classes of neurons with activity related to PGO have been discovered in PB and the

466

laterodorsal tegmental (LDT) nuclei (Steriade et al. 1990). Steriade et al. (1990) observed burst activity in several PGO-related cells in response to handclap stimuli which simultaneously elicited PGO E in LGB. This is the clearest evidence to date that P G O and P G O E have common neural origins. Whether CL and CL E are generated by similar cellular activity in the pons is to be determined. The observation of putative CNS markers of alerting (PGO E and CL E) without accompanying O R a n d / o r ASR indicates a degree of separation in the neural mechanisms underlying CNS responses from those governing peripheral responses to alerting stimuli. As expected, O R occurred infrequently across blocks in W and did not occur in N R E M and REM. However, we have found preliminary evidence that stimuli can elicit O R during R E M if atonia is eliminated by certain of the bilateral pontine tegmental lesions that release behavior in R E M (Morrison et al. 1990). We have also observed ASR in cats with pontine lesions (Morrison et al. 1990). In the present study, 3 non-lesioned cats exhibited ASR in W and in REM. The remaining 4 cats showed either sporadic, or no ASR in W and no ASR in REM. P G O E and CL E were obtained in all behavioral states. Not surprisingly, our findings for overt behaviors (OR and ASR) and evoked wave forms (PGO E and CL E) indicate that the CNS continues to process potentially relevant stimuli in the absence of motor responses. A limitation of this work is that it does not address the relationship between P G O E, CL E and directional eye movements. The E O G electrodes we use do not allow the determination of eye movement direction; and because our cats are freely moving we were unable to reliably observe eye movements behaviorally. The relationship between P G O and eye movement is well documented (e.g., Brooks 1967b; Brooks and Gershon 1971; Nelson et al. 1983), and eye movement may accompany P G O E (e.g., Bowker and Morrison 1977). Relevant to the issue of OR, Bowker and Morrison (1977) found that P G O E elicited by auditory or tactile stimuli in W preceded eye movement, whereas PGO E elicited by visual stimuli occurred after eye movement. Nelson et al. (1983) found that P G O amplitude in R E M was greater in the LGB ipsilateral to eye movement direction compared to that in the contralateral LGB. No one has examined the relationship of waves in CL to eye movements, although cells in the thalamic internal medullary lamina (including CL) appear to have an important role in the control of visually elicited eye movements (Schlag-Rey and Schlag 1977). A determination of the relationship between PGOE, CL E and directional eye movements could lend important insights into the neural mechanisms underlying OR. The original hypothesis that P G O and ASR are produced by the s a m e neural mechanisms (Bowker and

L.D. S A N F O R D E T AL.

Morrison 1976) has been modified in line with several lines of evidence (Kaufman 1983; Glenn 1985; Wu et al. 1989), which are supported by the present results. PGO E and motor responses respond differently to tones of different intensities, and across behavioral states (Wu et al. 1989; Ball et al. 1991a). At low stimulus intensities in W and NREM, the cat motor response is characterized by suppressed E M G activity, whereas at higher stimulus intensities an excitatory component is overlaid on the E M G suppression (Wu et al. 1989). In contrast, P G O E amplitude increases as a function of intensity and also varies as a function of state with the highest amplitudes found in R E M along with muscle atonia (Wu et al. 1989; Ball et al. 1991a). Wu et al. (1989) have argued that ASR, E M G suppression and P G O E reflect the operation of 3 independent phasic response systems. These systems may be activated concurrently by intense auditory stimuli in W, yet may respond independently to changes in stimulus intensity and behavioral state. In our tone habituation studies we have generally found a linear decrease in ASR amplitude across stimulus presentations in W (Sanford et al. 1992). Three of the cats in this study showed high amplitude ASR throughout the experiment with little or no sign of decrement in amplitude across blocks despite the fact that they had experienced at least 150 tones in previous experiments. These 3 cats did show a decrement in ASR amplitude across trials within blocks in W, suggesting that some level of dishabituation occurred in the 20 rain interval between blocks. In the remaining 4 cats ASR apparently had habituated. This diversity among individual animals is not surprising since ASR in cats appears to be inconsistent compared to the virtually ubiquitous ASR reported for rats (Davis 1985; Cassella and Davis 1986) and more difficult to elicit as measured by hindlimb E M G (Gruner 1989). PGO E were observed in all cats tested, an observation that supports the contention that motor and central components of alerting responses are controlled by separate neural mechanisms. The separation of P G O and ASR mechanisms may not actually be so clear. We found that ASR, as measured by an accelerometer, could be obtained along with P G O E in R E M even when no tonus was apparent in neck EMG. This finding contrasts with that of Wu et al. (1989) who recorded ASR in neck E M G in N R E M but not in REM. However, these disparate findings for ASR were based on two different recording techniques, and we did not find evidence of ASR in our recordings of neck E M G during REM. The characteristics of ASR in R E M matched the recordings of whole body ASR obtained in W and were distinct from the phasic muscle twitches normally seen in R E M and which were also apparent in the accelerometer record. Interestingly, the amplitude of ASR in 2 of these cats

ALERTING IN SLEEP AND WAKEFULNESS

did not appear to decrease across the first 40 tone presentations in REM as it did in W. A third cat that exhibited sporadic ASR in W only startled during the initial portion of the REM episode. Whether the difference between W and REM was due to state-dependent habituation processes or to a different overall level of excitability is not known. Previously, we suggested a tentative relationship between CL E and PGOE, based on data obtained only in W, in which the two waves mark the operation of neural mechanisms working to direct and enhance visual information gathering processes while at the same time preparing the cortex to receive that information (Ball et al. 1989a; Sanford et al. 1992). That is, in W, while the cortex is readied (desynchronized via the CL (e.g., Morison and Dempsey 1942; Steriade 1981; Glenn and Steriade 1982) to accept incoming information, the operation of a component of the visual system (e.g., the LGB) is being modified, perhaps, to sharpen that information (Singer 1973, 1977; Bowker and Morrison 1977). Our present results in REM and N R E M fit well with this view and suggest that these information gathering processes are operative in all states, although overt manifestation of behavioral responses (e.g., OR and ASR) may be reduced or absent. Whether this separation reflects the operation of common neural mechanisms that act to integrate a n d / o r disintegrate peripheral and central alerting responses, or whether a constellation of independent mechanisms involved in alerting exists that may be activated simultaneously in some instances and not in others remains to be determined.

References Ball, W.A., Morrison, A.R., Hunt, W. and Ross, R.J. Elicited PGO waves in slow-wave sleep and paradoxical sleep. In: J. Horne (Ed.), Sleep '88. Gustav Fischer, New York, 1989a: 207-209. Ball, W.A., Morrison, A.R. and Ross, R.J. The effects of tones of PGO waves in slow wave sleep and paradoxical sleep. Exp. Neurol., 1989b, 104: 251-256. Ball, W.A., Hunt, W., Sanford, L.D., Ross, R.J. and Morrison, A.R. Effects of stimulus intensity on elicited ponto-geniculo-occipital waves in sleep. Electroenceph. clin. Neurophysiol., 1991a, 78: 35-39 Ball, W.A., Sanford, L.D., Morrison, A.R., Ross, R.J., Hunt, W.K. and Mann, G . L The effects of changing state on elicited pontogeniculo-occipital (PGO) waves. Electroenceph. clin. Neurophysiol., 1991b, 79: 420-429. Berman, A.L. The Brainstem of the Cat. A Cytoarchitectonic Atlas with Stereotaxic Coordinates. University of Wisconsin Press, Madison, Wl, 1968. Bowker, R.M. and Morrison, A.R. The startle reflex and PGO spikes. Brain Res., 1976, 102: 185-190. Bowker, R.M. and Morrison, A.R. The PGO spike: an indicator of hyperalertness. In: W.P. Koella and P. Levin (Eds.), Sleep Research-1976. Karger, Basel, 1977: 23-27.

467 Brooks, D.C. Localization of the lateral geniculate nucleus waves associated with paradoxical sleep in the cat. Electroenceph. clin. Neurophysiol., 1967a, 23: 123-133. Brooks, D.C. Waves associated with eye movement in the awake and sleeping cat. Electroenceph. clin. Neurophysiol., 1967b, 24: 532541. Brooks, D.C. Localization and characteristics of the cortical waves associated with eye movement in the cat. Exp. Neurol., 1968, 22: 603-613. Brooks, D.C. and Gershon, M.D. Eye movement potentials in the occulomotor and visual systems of the cat: a comparison of reserpine induced waves with those present during wakefulness and rapid eye movement sleep. Brain Res., 1971, 27: 223-239. Cairo, J.M. and Fernandez-Guardiola, A. Phasic activity of the basolateral amygdala, cingulate gyrus and hippocampus during REM sleep in the cat. Sleep, 1984, 7: 202-210. Cassella, J.V. and Davis, M. The design and calibration of a startle measurement system. Physiol. Behav., 1986, 36: 377-383. Davis, M. The mammalian startle response. In: R. Eaton (Ed.), The Neural Basis of Startle Behaviour. Plenum Press, New York, 1985: 287-351. De Lima, A.D. and Singer, W. The brainstem projection to the lateral geniculate nucleus in the cat: identification of cholinergic and monoaminergic elements. J. Comp. Neurol., 1987, 259: 92121. Glenn, L.L Brainstem and spinal control of lower limb motorneurons with special reference to phasic events and startle reflexes. In: D.J. McGinty, R. Drucker-Colin, A.R. Morrison and P.L. Parmeggiani (Eds.), Brain Mechanisms of Sleep. Raven Press, New York, 1985: 81-95. Glenn, L.L. and Steriade, M. Discharge rate and excitability of cortically projecting intralaminar thalamic neurons during waking and sleep states. J. Neurosci., 1982, 2: 1387-1404. Gruner, J.A. Comparison of vestibular and auditory startle responses in the rat and cat. J. Neurosci. Meth., 1989, 27: 13-23. Hendricks, J.C., Morrison, A.R. and Mann, G.L. Different behaviors during paradoxical sleep without atonia depend on pontine lesion site. Brain Res., 1982, 239: 81-105. Henley, K. and Morrison, A.R. A re-evaluation of the effects of lesions of the positive tegmentum and locus coeruleus on phenomena of paradoxical sleep in the cat. Acta Neurobiol. Exp., 1974, 34: 215-232. Hobson, J.A. L'activit6 61ectrique phasique du cortex et du thalamus au cours du sommeil dgsynchronis6 chez le chat. C.R. Soc. Biol., 1964, 158: 2131-2134. Hu, B., Steriade, M. and Desch~nes, M. The effects of peribrachial stimulation on neurons of the lateral geniculate nucleus. Neuroscience, 1989a, 31: 13-24. Hu, B., Steriade, M. and Desch~nes, M. The cellular mechanisms of thalamic ponto-geniculo-occipital waves. Neuroscience, 1989b, 31: 25-35. Jeannerod, M. Organisation de l'activitg glectrique phasique du sommeil paradoxal. Etude 61ectrophysiologique et neuropharmacologique. Theses de m6decine, Lyon, 1965. Jeannerod, M. and Sakai, K. Occipital and geniculate potentials related to eye movements in the unanesthetized cat. Brain Res., 1970, 19: 361-377. Jouvet, M. Neurophysiology of the states of sleep. Physiol. Rev., 1967, 47: 117-177. Jouvet, M. and Delorme, F. Locus coeruleus et sommeil paradoxal. C.R. Soc. Biol., 1965, 159: 895-899. Kaufman, L.S. PGO waves in rats in the non-paradoxical sleep states. Brain Res., 1983, 276: 73-80. Marks, G.A., Farber, G. and Roffwarg, H.P. Metencephalic localization of ponto-geniculo-occipital waves in the albino rat. Exp. Neurol., 1980, 69: 667-677. McCarley, R.W., Nelson, J.P. and Hobson, J.A. Ponto-geniculo-oc-

468 cipital (PGO) burst neurons: correlative evidence for neuronal generator of PGO waves. Science, 1978, 201: 269-272. Morison, R.S. and Dempsey, E.W. A study of thalamo-cortical relations. Am. J. Physiol., 1942, 135: 281-292. Morrison, A.R. Brainstem regulation of behavior during sleep and wakefulness. In: J.M. Sprague and A.W. Epstein (Eds.), Psychobiology and Physiological Psychology. Academic Press, New York, 1979: 91-131. Morrison, A.R., Ball, W.A., Sanford, L.D., Mann, G.L. and Ross, R.J. Stimuli elicit behavior in paradoxical sleep without atonia. In: J. Horne (Ed.), Sleep '90. Pontenagle Press, Bochum, 1991: 90-93. Nelson, J.P., McCarley, R.W. and Hobson, J.A. REM sleep burst neurons, PGO waves, and eye movement information. J. Neurophysiol., 1983, 50: 784-797. Pare, D., Smith, Y., Parent, A. and Steriade, M. Projections of brainstem core cholinergic and noncholinergic neurons of cat to intralaminar and reticular thalamic nuclei. Neuroscience, 1988, 25: 69-86. Pellet, J., Tardy, M., Dubrocard, S. and Harley, R. Etude de l'activit~ 61ectrique phasique du cortex du cervelet au cours des ~tats de veille et de sommeil. Arch. Ital. Biol., 1974, 112: 163-195. Sakai, K. and Jouvet, M. Brainstem PGO - on cells projecting directly to the cat dorsal lateral geniculate nucleus. Brain Res., 1980, 194: 500-505. Sanford, L.D., Ball, W.A., Morrison, A.R., Ross, R.J. and Mann, G. Peripheral and central components of alerting: habituation of acoustic startle, orienting responses and elicited waveforms. Behav. Neurosci., 1992, 106: 1-8. Sastre, J. and Jouvet, M. Le comportement on6'/rique du chat. Physiol. Behav., 1979, 22: 979-989. Schlag-Rey, M. and Schlag, J. Visual and presaccadic neuronal activity in thalamic internal medullary lamina of cat: a study of targeting. J. Neurophysiol., 1977, 40: 156-173. Singer, W. The effect of mesencephalic reticular stimulation on intracellular potentials of cat lateral geniculate neurons. Brain Res., 1973, 61- 35-54. Singer, W. Control of thalamic transmission by corticofugal and ascending reticular pathways in the visual system. Physiol. Rev., 1977, 57: 386-420. Smith, Y., Pare, D., Desch~nes, M., Parent, A. and Steriade, M.

L.D. SANFORD ET AL. Cholinergic and non-cholinergic projections from the brainstem and ports to the visual thalamus in the cat. Exp. Brain Res., 1988, 70: 166-180. Sofroniew, M.V., Priestley, J.V., Consolazione, A., Echenstein, F. and Cuello, C.A. Cholinergic projections from the midbrain and pons to the thalamus in the rat, identified by combined retrograde tracing and choline acetyltransferase immunohistochemistry. Brain Res., 1985, 329: 213-223. Sokolov, E.N. The modeling properties of the nervous system. In: M. Cole and I. Maltzman (Eds.), A Handbook of Contemporary Soviet Psychology. Basic Books, New York, 1969: 671-704. Sprague, J.M., Levitt, M., Robson, K., Liu, C.N., Stellar, E. and Chambers, W.W. A neuroanatomical and behavioral analysis of the syndromes resulting from midbrain lemniscal and reticular lesions in the cat. Arch. Ital. Biol., 1963, 101: 225-295. Steriade, M. Mechanisms underlying cortical activation: neuronal organization and properties of the midbrain reticular core and intralaminar thalamic nuclei. In: O. Pompeiano and C. Ajmone Marsan (Eds.), Brain Mechanisms and Perceptual Awareness. Raven Press, New York, 1981: 327-377. Steriade, M. and Glenn, L.L. Neocortical and caudate projections of intralaminar neurons and their synaptic excitation from midbrain reticular core. J. Neurophysiol., 1982, 48: 352-371. Steriade, M., Pare, D., Parent, A. and Smith, J.C. Projections of cholinergic and non-cholinergic neurons of the brainstem core to relay and associational thalamic nuclei in the cat and macaque monkey. Neuroscience, 1988, 25: 47-67. Steriade, M., Pare, D., Datta, S., Oakson, G. and Curro Dossi, R. Different cellular types in mesopontine cholinergic nuclei related to ponto-geniculo-occipital waves. J. Neurosci., 1990, 10: 25462565. Ursin, H., Sundberg, H. and Menaker, S. Habituation of the orienting response elicited by stimulation of the caudate nucleus in the cat. Neuropsychology, 1969, 7: 313-318. Woolf, N.J., Harrison, J.B. and Buchwald, J.S. Cholinergic neurons of the feline pontomesencephalon. If. Ascending anatomical projections. Brain Res., 1990, 520: 55-72. Wu, M.F., Mallick, B.N. and Siegel, J.M. Lateral genieulate spikes, muscle atonia and startle response elicited by auditory stimulus parameters and arousal state. Brain Res., 1989, 499: 7-18.