Dissociated pattern of activity in visual cortices and their projections during human rapid eye movement sleep

Science

Volumn 279, Issue 5347, 1998, Pages 91-95

  Braun, Allen R ^a   Balkin, Thomas J ^a   Wesensten, Nancy J ^a   Gwadry, Fuad ^a   Carson, Richard E ^a   Varga, Mary ^a   Baldwin, Paul ^a   Belenky, Gregory ^a   Herscovitch, Peter ^a  

^a NONE

Author keywords

[No Author keywords available]

Indexed keywords

OXYGEN 15;

ARTICLE; BRAIN BLOOD FLOW; BRAIN CORTEX; HUMAN; HUMAN EXPERIMENT; MALE; NORMAL HUMAN; POSITRON EMISSION TOMOGRAPHY; PRIORITY JOURNAL; REM SLEEP; SLEEP DEPRIVATION; SLEEP WAKING CYCLE; VISUAL CORTEX; VISUAL NERVOUS SYSTEM;

ADULT; BRAIN MAPPING; DREAMS; HIPPOCAMPUS; HUMANS; LIMBIC SYSTEM; MALE; PREFRONTAL CORTEX; REGIONAL BLOOD FLOW; SLEEP; SLEEP, REM; TOMOGRAPHY, EMISSION-COMPUTED; VISUAL CORTEX; VISUAL PATHWAYS; WAKEFULNESS;

EID: 0032472267     PISSN: 00368075     EISSN: None     Source Type: Journal    
DOI: 10.1126/science.279.5347.91     Document Type: Article

References (52)

1. E. Aserinsky and N. Kleitman, Science 118, 273 (1953).
2. W. Dement and N. Kleitman, J. Exp. Psychol. 53, 339 (1957).
3. Some studies suggest that activation of only the higher levels of the visual processing hierarchy is sufficient to evoke and sustain visual imagery [P. E. Roland and B. Gulyas, Cereb. Cortex 5, 79 (1995)].
4. Others [S. M. Kosslyn, W. L. Thompson, I. J. Kim, N. M. Alpert, Nature 378, 496 (1995)] indicate that contribution of the primary visual processing areas - possibly activated through reentrant pathways [D. J. Felleman and D. C. Van Essen, Cereb. Cortex 1, 1 (1991)] - is also required.
5. Others [S. M. Kosslyn, W. L. Thompson, I. J. Kim, N. M. Alpert, Nature 378, 496 (1995)] indicate that contribution of the primary visual processing areas - possibly activated through reentrant pathways [D. J. Felleman and D. C. Van Essen, Cereb. Cortex 1, 1 (1991)] - is also required.
6. J. A. Hobson, The Dreaming Brain (Basic Books, New York, 1988).
7. Informed consent was obtained from 10 healthy male volunteers (23.6 ± 1.7 years of age, mean ± SD; range, 22 to 27 years). Eight subjects were right-handed, and two were left-handed.
8. There were 8 hours of sleep with selective REM deprivation on day 1 ; 2 hours of sleep with REM deprivation on day 2; and 15.5 to 23.5 hours of continuous wakefulness (22.2 ± 2.4 hours, mean ± SD) on day 3. Subjects were monitored polysomnographically throughout this period to verify the absence of sleep and facilitate selective REM deprivation.
9. PET scans were performed on a Scanditronix PC2048-15B tomograph (Uppsala, Sweden), which acquires 15 contiguous planes. A transmission scan was performed for attenuation correction.
10. 15O were injected and scans were initiated automatically. For inclusion in these analyses, a sleep stage had to be uniformly maintained (free of arousals and stage shifts) from 40 s before to 40 s after scan initiation. It was necessary to use partial sleep-deprivation methods to help ensure that subjects would sleep in the scanner. Such procedures, however, may affect sleep architecture (that is, the relative amounts and timing of the various stages within the sleep period). To evaluate these potential effects, we compared our subjects' sleep architecture with published normative data [R. L. Williams, I. Karacan, C. J. Hursch, Electroencephalography (EEG) of Human Sleep: Clinical Applications (Wiley, New York, 1974)]. Except for increased SWS time during the first third of the sleep period (an expected consequence of the deprivation procedure), we found no qualitative or quantitative abnormalities in the sleep architecture of our subjects. Furthermore, the scans selected for analysis were, in each instance, associated with sleep samples that conformed explicitly and unambiguously to the standard scoring criteria of Rechtschaffen and Kales throughout the period of data acquisition. Therefore, although sleep deprivation may have affected duration and timing of sleep stages, the procedure produced no observable effects on the physiological characteristics of the sleep stages themselves (that is, at the level of analysis most critical to this study).
11. 15O were injected and scans were initiated automatically. For inclusion in these analyses, a sleep stage had to be uniformly maintained (free of arousals and stage shifts) from 40 s before to 40 s after scan initiation. It was necessary to use partial sleep-deprivation methods to help ensure that subjects would sleep in the scanner. Such procedures, however, may affect sleep architecture (that is, the relative amounts and timing of the various stages within the sleep period). To evaluate these potential effects, we compared our subjects' sleep architecture with published normative data [R. L. Williams, I. Karacan, C. J. Hursch, Electroencephalography (EEG) of Human Sleep: Clinical Applications (Wiley, New York, 1974)]. Except for increased SWS time during the first third of the sleep period (an expected consequence of the deprivation procedure), we found no qualitative or quantitative abnormalities in the sleep architecture of our subjects. Furthermore, the scans selected for analysis were, in each instance, associated with sleep samples that conformed explicitly and unambiguously to the standard scoring criteria of Rechtschaffen and Kales throughout the period of data acquisition. Therefore, although sleep deprivation may have affected duration and timing of sleep stages, the procedure produced no observable effects on the physiological characteristics of the sleep stages themselves (that is, at the level of analysis most critical to this study).
12. 2 values. Of the 10 subjects in whom REM sleep scans were acquired, 7 had SWS scans, and 7 had waking postsleep studies. Stage contrasts were performed in a repeated measures design; thus, N = 7 for both REM-waking and REM-SWS contrasts. To evaluate hemispheral effects, differences in normalized rCBF for each contrast were compared between homologous pixels in right and left hemispheres by using a voxel-wise error variance; differences exceeding a Z-threshold of 3.09 were considered significant.
13. 2 values. Of the 10 subjects in whom REM sleep scans were acquired, 7 had SWS scans, and 7 had waking postsleep studies. Stage contrasts were performed in a repeated measures design; thus, N = 7 for both REM-waking and REM-SWS contrasts. To evaluate hemispheral effects, differences in normalized rCBF for each contrast were compared between homologous pixels in right and left hemispheres by using a voxel-wise error variance; differences exceeding a Z-threshold of 3.09 were considered significant.
14. 2 values. Of the 10 subjects in whom REM sleep scans were acquired, 7 had SWS scans, and 7 had waking postsleep studies. Stage contrasts were performed in a repeated measures design; thus, N = 7 for both REM-waking and REM-SWS contrasts. To evaluate hemispheral effects, differences in normalized rCBF for each contrast were compared between homologous pixels in right and left hemispheres by using a voxel-wise error variance; differences exceeding a Z-threshold of 3.09 were considered significant.
15. The frequency of rapid conjugate eye movements, identified as positive or negative deflections on the electrooculogram (EOG) during the 180 s after the start of the scan, were computed. Normalized rCBF images were correlated with each subject's REM counts on a pixel-by-pixel basis (N = 10 for these analyses) (Fig. 1C). Correlation coefficients were thresholded at a level of 0.6 (P < 0.05, n = 10) for tabulation (Table 1).
16. Coordinates for 20 visual cortical areas in each hemisphere (7 striate, 6 lateral occipital, 7 fusiform/infero-temporal) were selected by using the Talairach atlas, adjusted with the SPM PET template, and used to extract rCBF values from the stereotaxically normalized PET images. These values were normalized by using mean CBF rates for the visual cortex as a whole; rCBF values in homologous regions of right and left hemispheres were averaged and used to generate difference matrices (REM-SWS and REM-wake; N = 7); a principal component analysis was then applied. Significance of the extracted component vector was assessed by using permutation tests [B. F. J. Manly, Randomization and Monte Carlo Methods in Biology (Chapman & Hall, New York, 1990)] by randomly permuting the sign of data entries 19 times; P values were derived by determining the proportion of permuted trials with higher first eigenvalues than the original data. Stability of regional loadings was assessed by the bootstrap procedure [B. Efron, SIAM Monogr. 38-1 (1982)]. Fit values of the original loadings across 10,000 bootstrap samples, expressed as Z-scores, were calculated; absolute values > 2.4 were considered stable.
17. Coordinates for 20 visual cortical areas in each hemisphere (7 striate, 6 lateral occipital, 7 fusiform/infero-temporal) were selected by using the Talairach atlas, adjusted with the SPM PET template, and used to extract rCBF values from the stereotaxically normalized PET images. These values were normalized by using mean CBF rates for the visual cortex as a whole; rCBF values in homologous regions of right and left hemispheres were averaged and used to generate difference matrices (REM-SWS and REM-wake; N = 7); a principal component analysis was then applied. Significance of the extracted component vector was assessed by using permutation tests [B. F. J. Manly, Randomization and Monte Carlo Methods in Biology (Chapman & Hall, New York, 1990)] by randomly permuting the sign of data entries 19 times; P values were derived by determining the proportion of permuted trials with higher first eigenvalues than the original data. Stability of regional loadings was assessed by the bootstrap procedure [B. Efron, SIAM Monogr. 38-1 (1982)]. Fit values of the original loadings across 10,000 bootstrap samples, expressed as Z-scores, were calculated; absolute values > 2.4 were considered stable.
18. -7; centered x = 21.9, y = -51.8, z = -6.8.
19. S. M. Kosslyn and K. N. Ochsner, Trends Neurosci. 17, 290 (1994).
20. A. Rechtschaffen, P. Verdone, J. Wheaton, Can. Psychiatr. Assoc. J. 8, 409 (1963).
21. -6; centered x = 2.8, y = -83.0, z = 0.4.
22. 2 corrected flow rates were consistently lower during REM sleep than during SWS throughout the striate cortex [ΔrCBF = -6.40 ± 5.9 ml/100 g/min (mean ± SD), x = 4, y = -82, z = 0], and absolute rCBF rates were consistently elevated in extrastriate regions (in the fusiform cortex, ΔrCBF = +8.85 ± 2.79 ml/100g/min, x = -34, y = -56, z = -8), although variances in absolute values were large.
23. max > k), corrected = 3.55E - 15; centered x = -3.9, y = 0.3, z = 1.2.
24. D. A. Chavis and D. N. Pandya, Trans. Am. Neurol. Assoc. 99, 192 (1974); W. A. Suzuki and D. G. Amaral, J. Comp. Neurol. 350, 497 (1994).
25. D. A. Chavis and D. N. Pandya, Trans. Am. Neurol. Assoc. 99, 192 (1974); W. A. Suzuki and D. G. Amaral, J. Comp. Neurol. 350, 497 (1994).
26. The frontal eye fields (superior dorsolateral prefrontal and premotor cortices) were incompletely sampled in many subjects and are not included in this analysis.
27. The bootstrapping procedures indicated that most regional loadings were stable. In the REM-SWS contrast, all the fusiform-interotemporal regions - but none of the lateral occipital regions - were associated with Z-scores exceeding threshold (local maximum = 5.44, x = ±48, y = -62, z = 0). In the REM-wake contrast, six of the seven fusiform-inferotemporal regions (maximum 4.18, x = ±48, y = -62, z = 4), and four of the six lateral occipital regions (maximum 3.58, x = ±32, y = -86, z = 0) were stable; in the latter case, all were located in ventral rather than dorsal portions of the lateral occipital cortex (z ≤ 8 mm). In the REM-SWS contrast, five of the seven striate loadings were stable (local minimum = -3.79, x = ±4, y = -82, z = 0), and in the REM-wake contrast, six of seven loadings were stable (minimum = -9.89, x = ±4, y = -82, z = 0).
28. S. Zela et al., J. Neurosci. 11, 641 (1991); D. Watson et al., Cereb. Cortex 3, 79 (1993); J. V. Haxby et al., J. Neurosci. 14, 6336 (1994); I. Sereno et al., Science 268, 889 (1995).
29. S. Zela et al., J. Neurosci. 11, 641 (1991); D. Watson et al., Cereb. Cortex 3, 79 (1993); J. V. Haxby et al., J. Neurosci. 14, 6336 (1994); I. Sereno et al., Science 268, 889 (1995).
30. S. Zela et al., J. Neurosci. 11, 641 (1991); D. Watson et al., Cereb. Cortex 3, 79 (1993); J. V. Haxby et al., J. Neurosci. 14, 6336 (1994); I. Sereno et al., Science 268, 889 (1995).
31. S. Zela et al., J. Neurosci. 11, 641 (1991); D. Watson et al., Cereb. Cortex 3, 79 (1993); J. V. Haxby et al., J. Neurosci. 14, 6336 (1994); I. Sereno et al., Science 268, 889 (1995).
32. These results could reflect the central effects of saccadic eye movements per se. Voluntary saccades during wakefulness previously have been found to be associated with reductions in visual CBF [T. Paus, S. Marrett, K. J. Worsley, A. C. Evans, J. Neurophysiol. 74, 2179 (1995)]. However, saccades in that study were continuous and rapid (40 to 140 in 60 s) but were intermittent and lower in frequency in our own study [0 to 11 in 60 s (5.5 ± 4.0, mean ± SD]. Furthermore, reductions in rCBF were more widely distributed throughout the visual cortex, and no positive correlations between extrastriate activity and saccadic eye movements were observed, which suggests that the striate-extrastriate dissociation evident in this study may be a unique characteristic of REM sleep.
33. R. W. McCarley, J. W. Winkelman, F. H. Duffy, Brain Res. 274, 359 (1983); C. W. Callaway, R. Lydic, H. A. Baghdoyan, J. A. Hobson, Cell Mol. Neurobiol. 7, 105 (1987).
34. R. W. McCarley, J. W. Winkelman, F. H. Duffy, Brain Res. 274, 359 (1983); C. W. Callaway, R. Lydic, H. A. Baghdoyan, J. A. Hobson, Cell Mol. Neurobiol. 7, 105 (1987).
35. The effects of the PGO wave generator on the occipital cortex do not appear to be mediated through the lateral geniculate nucleus [J. A. Hobson, J. Alexander, C. J. Frederickson, Brain Res. 14, 607 (1969)] and therefore are not expected to be exclusively associated with striate projections. Early studies also demonstrated that PGO wave amplitudes in the lateral association areas of the cat visual cortex exceed those in primary visual cortices [D. C. Brooks, Exp. Neurol. 22, 603 (1968)], and single unit studies have shown that the facilitatory effects of PGO waves on neuronal firing rates appear to be manifest in extrastriate but not in striate cortices [T. Kasamatsu and W. R. Adey, Brain Res. 55, 323 (1973)].
36. The effects of the PGO wave generator on the occipital cortex do not appear to be mediated through the lateral geniculate nucleus [J. A. Hobson, J. Alexander, C. J. Frederickson, Brain Res. 14, 607 (1969)] and therefore are not expected to be exclusively associated with striate projections. Early studies also demonstrated that PGO wave amplitudes in the lateral association areas of the cat visual cortex exceed those in primary visual cortices [D. C. Brooks, Exp. Neurol. 22, 603 (1968)], and single unit studies have shown that the facilitatory effects of PGO waves on neuronal firing rates appear to be manifest in extrastriate but not in striate cortices [T. Kasamatsu and W. R. Adey, Brain Res. 55, 323 (1973)].
37. The effects of the PGO wave generator on the occipital cortex do not appear to be mediated through the lateral geniculate nucleus [J. A. Hobson, J. Alexander, C. J. Frederickson, Brain Res. 14, 607 (1969)] and therefore are not expected to be exclusively associated with striate projections. Early studies also demonstrated that PGO wave amplitudes in the lateral association areas of the cat visual cortex exceed those in primary visual cortices [D. C. Brooks, Exp. Neurol. 22, 603 (1968)], and single unit studies have shown that the facilitatory effects of PGO waves on neuronal firing rates appear to be manifest in extrastriate but not in striate cortices [T. Kasamatsu and W. R. Adey, Brain Res. 55, 323 (1973)].
38. D. R. Goodenough, H. B. Lewis, A. Shapiro, I. Sleser, J. Nerv. Ment. Dis. 140, 365 (1965); P. Verdone, Percept. Mot. Skills 20, 1253 (1965); T. Pivik and D. Foulkes, Science 153, 1282 (1966); W. Dement and E. A. Wolpert, J. Exp. Psychol. 55, 543 (1958). See also (2, 14).
39. D. R. Goodenough, H. B. Lewis, A. Shapiro, I. Sleser, J. Nerv. Ment. Dis. 140, 365 (1965); P. Verdone, Percept. Mot. Skills 20, 1253 (1965); T. Pivik and D. Foulkes, Science 153, 1282 (1966); W. Dement and E. A. Wolpert, J. Exp. Psychol. 55, 543 (1958). See also (2, 14).
40. D. R. Goodenough, H. B. Lewis, A. Shapiro, I. Sleser, J. Nerv. Ment. Dis. 140, 365 (1965); P. Verdone, Percept. Mot. Skills 20, 1253 (1965); T. Pivik and D. Foulkes, Science 153, 1282 (1966); W. Dement and E. A. Wolpert, J. Exp. Psychol. 55, 543 (1958). See also (2, 14).
41. D. R. Goodenough, H. B. Lewis, A. Shapiro, I. Sleser, J. Nerv. Ment. Dis. 140, 365 (1965); P. Verdone, Percept. Mot. Skills 20, 1253 (1965); T. Pivik and D. Foulkes, Science 153, 1282 (1966); W. Dement and E. A. Wolpert, J. Exp. Psychol. 55, 543 (1958). See also (2, 14).
42. This appears to be consistent with the finding of Roland and Gulyas (3). However, our results may not be directly comparable with studies of conscious visual imagery; task-elicited activation of the primary visual cortex seen by Kosslyn et al. (4) might be a special feature of conscious visual imagery or retrieval processes that occur during the waking state.
43. P. Maquet et al. [Nature 383, 163 (1996)] did not list rCBF changes in visual cortices; however, they compared REM sleep with an amalgam of wake and non-REM stages, whereas our contrasts were stage specific (for example, REM-SWS) and may be more sensitive.
44. P. L. Madsen et al. [J. Cereb. Blood Flow Metab. 11, 502 (1991)] reported REM-associated increases in rCBF in visual association cortices; functional relationships between activity in striate and extrastriate cortices were not evaluated.
45. C. C. Hong et al. [Sleep 18, 570 (1995)] reported a positive correlation between REM density and glucose metabolic rates in lateral occipital cortex, but results for primary visual cortex were not reported.
46. S. Zeki, A Vision of the Brain (Blackwell, Oxford, 1993). See also (3, 4).
47. A pattern of extrastriate activity in the absence of striate function also characterizes clinical states associated with unusual and often bizarre perturbations of visual awareness, such as blindsight [L. Weiskrantz, J. L. Barbur, A. Sahraie, Proc. Natl. Acad. Sci. U.S.A. 92, 6122 (1995)] and confabulatory denial of blindness [G. Goldenberg, W. Mullbacher, A. Nowak, Neuropsychologia 33, 1373 (1995)]. The features of both syndromes suggest that synchronous activation of primary and extrastriate cortices may be essential for "normal" visual awareness. That such coherence appears to be breached during REM sleep is not inconsistent with the altered awareness that characterizes this sleep stage.
48. A pattern of extrastriate activity in the absence of striate function also characterizes clinical states associated with unusual and often bizarre perturbations of visual awareness, such as blindsight [L. Weiskrantz, J. L. Barbur, A. Sahraie, Proc. Natl. Acad. Sci. U.S.A. 92, 6122 (1995)] and confabulatory denial of blindness [G. Goldenberg, W. Mullbacher, A. Nowak, Neuropsychologia 33, 1373 (1995)]. The features of both syndromes suggest that synchronous activation of primary and extrastriate cortices may be essential for "normal" visual awareness. That such coherence appears to be breached during REM sleep is not inconsistent with the altered awareness that characterizes this sleep stage.
49. Decreases in activity in the prefrontal cortex during REM sleep have been noted previously by Madsen et al. (29) and by Maquet et al. (28). On the other hand, Hong et al. (30) reported positive correlations between REM density and absolute glucose metabolic rates in the (right) dorsolateral prefrontal cortex; however, this measure, which reflects physiological responses over a longer time period, may not be directly comparable with normalized rCBF values evaluated in this study.
50. A. Rechtschaffen, Sleep 1, 97 (1978); J. A. Hobson, Endeavour 20, 86 (1996).
51. A. Rechtschaffen, Sleep 1, 97 (1978); J. A. Hobson, Endeavour 20, 86 (1996).
52. The authors wish to thank Dr. Alex Martin for his expertise, insight, and critical suggestions, which were essential in the preparation of this manuscript, and Dr. Scott Selbie for his valuable help in preparing Fig. 1.