Study of waking-sleeping behaviour using

STUDY OF WAKING-SLEEPING BEHAVIOUR USING AUTOMATIC ANALYSIS AND QUANTIFICATION CI. Gottesmann, G. Lacoste, L. Rodrigues, P. Kirkham, J.L. Rallo and Ch. Arnaud Laboratoire de Psychophysiologie. Facuit# des Sciences et des Techniques 06034 NICE Cedex

The activation of nerve and muscle cells is characterised by a iow voltage a c t i v i t y . The study of these fluctuations of potential difference is the purpose of the electrophysiological approach widely used in c l i n i c a l medecine on one hand, psychophysioiogicai and neurophysiological research on the other.

This e l e c t r i c a l a c t i v i t y has e s s e n t i a l l y two forms : - the f i r s t is rapid, of digital type, since i t is either on or off; this is what one observes in the axons, the nervous conducting paths, and single cell recordings : i t is a matter of calculating a response frequency.

-

the second consists of variations of potential difference with variable polarity, certain of them can last more than a second. They generally come from complex structures usually studied through several different layers of tissue. The typical example is that of the electroencephalogram recorded on the scalp or the electrocorticogram recorded direct l y on the cortex, the frequency of which varies from about 0.5-40 c/s. The study of this kind of a c t i v i t y is essential in Man for c l i n i c a l medecine, in animals for fundamental and applied research.

!32 This approach meets two major d i f f i c u l t i e s

:

- the mass of information from which the c r i t i c a l characteristics are to be drawn. -

visual data analysis, which is slow and open to inconsistency, This is why many research workers have set out to find automatic

analysis methods, which are both rapid and accurate, For our part, we have undertaken the automatic analysis of the d i f f e r e n t phases of waking and sleeping in the r a t , fundamental behaviour on which a l l else is based. This approach has been complemented by automatic analysis of brain e x c i t a b i l i t y and psychomotor a c t i v i t y . This research f i r s t l y

carried out o f f - l i n e , subsequently on-line, w i l l very

soon r e s u l t in a system allowing the automatic analysis of the same data on animals completly free to move. Seven phases of the waking-sleeping cycle have been selected, t h e i r physiological significance has been specified in our previous works ( i , 2) : 1/ active or a t t e n t i v e waking, with theta a c t i v i t y ; 2/ normal waking; 3/ slow wave sleep; 4/ deeper sleep characterised by neocortical spindles; 5/ intermediate stage which precedes and follows paradoxical sleep and is characterised by neocortical spindles and theta a c t i v i t y ; 6/ paradoxical sleep, where dreams occur in Man; 7/ the periods of rapid eye movements during paradoxical sleep (figure 1), In order to avoid any loss of information, we have set the unit duration of analysis at one second, I t is well known that the nervous system is in permanent unstable equilibrium, even more so in a small animal with higher metabolic rate, having a very polyphasic wakingsleeping rhythm with b r i e f a c t i v i t i e s such as spindles and the bursts of eye-movements, Off~line approach A/ State determination { f o r references, see 3~ 4, 5, 6, 7),

133 A study of the two cerebral derivations used, has shown that signal analysis by autocorreiation or cross correlation would require computing power far beyond the means available to us and for the pro-. cessing i t would be d i f f i c u l t

to reconcile signal length with the one

second' "quantum" we had decided on, Spectral analysis of the two brain derived channels has indicated that in a p a r t i c u l a r frequency band, the amount of energy in channel i , makes i t possible to distinguish between the principal sleeping and waking phases, with some overlapping. For channel 2, use is made of the r a t i o of energies in two frequency bands one of which is centred on the theta type a c t i v i t y . For the electrooculogram (channel 3) which is only taken into consideration during paradoxical sleep, the deviations from the base l i n e are detected, then integrated. F i n a l l y , for the electromyogram (channel 4), the energy emitted in a p a r t i c u l a r frequency band is integrated; this too d i f f e r s during the d i f f e r e n t behavioural phases, with overlaps, Thus each of the 3 essential derivations (1, 2, 4) is tested a certain number of times per unit of time; the energy collected is integrated over the one second base, This results in an energy value, for each derivation ~ which is specific to each phase of waking-sleeping behaviour with overlaps at the edges and which constitutes a p a r t i a l criterion, Each behavioural phase is also defined by a global c r i t e r i o n resulting from the combination of the three p a r t i a l c r i t e r i a , However, in certain cases one of the p a r t i a l c r i t e r i a is below the experimentally fixed norm, for the corresponding state, This results in a penalisation of the global criterion, which is a function of the unsatisfactory c r i terion:s deviation from the norm, Above a certain degree of penalisation, the state is retained but qualified as "doubtful", The doubtful states are counted with their probable state and also, separately, They are thus easely deductible, This analysis is undertaken on an IBM 1800 with 16 K words of 16 bits which processes the data at t h i r t y - t w o times the speed of acquisltion~ which is done in analogue form on magnetic tape, The signals undergo a pre~processing by means of pass~band f i l t e r s centred

134 on zones determined by spectral analysis. Each channel is sampled at 25-50 Hz depending on the derivation. The data are then d i g i t i s e d , Different rejection logics were used : for example, to eliminate heart pulses, high energy signals which are surimposed on the electromyogram; also transient states l i k e the short periods of clonus a c t i v i t i e s during paradoxical sleep that could be taken for attentive waking. Output : Cards are punched with the successive states and t h e i r duration, in binary code. The doubtful states are given in decreasing order of t h e i r p r o b a b i l i t y of appearance. The processing condenses 24 hours of recording into a maximum of 300 cards, These data are then grouped by quarter of an hour, hour and total length of the recording. The values produced in a l i s t i n g (figure 2A)show the percentage of time passed in each of the seven states, in waking as a whole, in sleeping as a whole, and the giobai percentage of doubtful states grouping together the individual percentages for each of the seven states~ F i n a l l y , an output in the form of histograms (figure 2B)indicates the evolution by quarter of an hour, of the percentage of time passed in each state and the average percentage for the whole recording, Reliability

:

The contingency studied by the C test (12) indicates the following values : Corrector A - Corrector A ~

C = ,89

Corrector B - Corrector B ;

C = ,83

Corrector A - Corrector B ;

C = .87

C o r r e c t o r A - Computer

;

C = ,81

C o r r e c t o r B ~ Computer

;

C = .81

135 The perfect contingency in our experimental conditions (global contingency between 6 states) is .91. The mean percentage of correspondence, state by state (13), between the two correctors and the computer is 82%. For the d i f f e r e n t stages i t is : attentive or active waking : 86% normal waking

: 79,5%

slow wave sleep

: 83%

spindles

: 78,5%

intermediate state

: 81%

paradoxical sleep

: 85,5%

These results are satisfactory. In f a c t , since the publication of our technique KOHN et a l . (14) have described a method with which they dissociate only three phases, with a 12 second period of integration and they obtain a 91% correspondence (15). B/ Det__er_mi_nation of_brain ex_cita__b~l~tZ I . Study of evoked potentials (for references, see 6, 7, 8, 9). A central or peripheral stimulation induces in the nervous pathways and centres with which i t comes into contact, an a c t i v i t y said to be "evoked", which gives indications about the level of e x c i t a b i l i t y of the central zone studied. The "evoked potential" is recorded in analogue form on a magnetic tape. During o f f - l i n e treatment, at 16 times the acquisition speed, a synchronising signal preceding the evoked potential opens an analysis window of variable duration with a high d e f i n i t i o n (2 samples per m i l l i second), during which the response is registered by the computer, then stored with the responses collected in the same behavioural phase. The responses which occur during "doubtful" states are eliminated. The mean amplitude and dispersion are automatically calculated for each sample point making up the evoked potential. Only a predetermined number of responses is retained, by hour and by state. The output of results takes place every twelve hours in graphical analogue form. The standard deviation is also marked on the same figure (~) -figure 3A-. 2

136

2. Analysis of induced states ( f o r references, see 6,7,9). Equally, a central or peripheral stimulation can also induce a change in behaviour and in the spontaneous global e l e c t r o p h y s i o l o g i c a l a c t i v i t i e s . A stimulus can momentarily wake the animal etc . . . This possible influence is tested by the automatic comparison of the animal's state in the second which precedes and the second which follows the stimulation . In order to avoid any interference with the state determination, the stimulus intended to induce an evoked p o t e n t i a l always occurs at the l i m i t of a period of state analysis ° The results of this analysis are produced as a l i s t i n g showing the percentage of induced states in each phase of the cycle and the number of samples counted ( f i g u r e 3B). C/ Psychomotor a c t i v i t y ( f o r references, see 7, 8, 9) . I t is important to know the rate and the d i s t r i b u t i o n of psychomotor a c t i v i t y during the circadian period in order to characterise the behaviour of a "normal" animal and that of an experimental animal subjected to various disturbances : s e l e c t i v e lesions of the nervous system or pharmacological influences . In order to detect movements of the preparation the electromyogram channel undergoes a second processing which allows the f l u c t u a t i o n s in energy in successive periods of analysis to be tested . Above an experimentally determined threshold, this f l u c t u a t i o n is counted as a movement, which is then c l a s s i f i e d with the state during which i t occurred . This approach has, in a d d i t i o n , the advantage of distinguishing the drugs which create a dissociation between the behaviour and the e l e c t r o corticogram, such as morphine and atropine which tend to induce a c o r t i c a l a c t i v i t y l i k e slow wave sleep although the animal has a waking behaviour and moves about im i t s cage . This diagnostic is established where an animal present a s i g n i f i c a n t increase in i t s movements during slow wave sleep, a phase which is usually associated with resting behaviour .

I37 The results are given as a l i s t i n g in the same way as f o r the states . The motor a c t i v i t y a r i s i n g in the d i f f e r e n t states is indicated in the form of a percentage of the duration of the state . On l i n e approach The o f f l i n e approach, rapid and rich in information, is interesting but i t nervertheless presents the following d i f f i c u l t i e s

-

:

i t is not always possible to detect experimental problems and artefacts during the data a c q u i s i t i o n ; the d i f f i c u l t i e s only appear a f t e r analysis.

- the determination of central e x c i t a b i l i t y requires the number of stimul a t i o n s to be increased, with the computer undertaking a sorting process afterwards . -

on a practical l e v e l , the analysis done outside the laboratory is very expensive, which has caused this technique to be abandoned . The purchase of a small computer (TEXAS 980A) has allowed us to

undertake the study of a program f o r analysis and q u a n t i f i c a t i o n in real time . Eliminating the d i f f i c u l t i e s outlined above, i t makes i t possible in addition to envisage a real-time i n v e s t i g a t i o n of metabolic changes . At the moment, we are studying the coupling of a push-pull cannula technique, f o r the perfusion of a central structure by means of 2 concent r i c or p a r a l l e l tubes, with the time basis of the computer . In this way, the output could be studied taking into account the precise state during which i t was obtained, and i t be easier to look f o r possible assimilation or e l i m i n a t i o n by the central nervous system, in terms of the d i f f e r e n t behavioural phases . I t is d i f f i c u l t

to indicate the precise method used as this

i n v e s t i g a t i o n is the subject of a contract . The real-time program has a better performance than the o f f - l i n e one and allows the feed-back necessary f o r the determination of central e x c i t a b i l i t y with a minimal number of stimulations .

The output data are the same as f o r the o f f - l i n e program . Figure 4 gives an example of recording analysed in real time without the i n t e r -

138 vention of an organisational l o g i c f o r the stages which would enabie the rare aberrant states to be eliminated. Real time processing of telemetry data Idea!]y one wishes to work on animal which are free to move. Up to now, we have only been able to use rats equipped with recording cables, which, though l i g h t , cause the animal discomfort and favour the dislodging of the a c r y l i c cement cap f i x e d to the s k u l l . Since a short while ago the laboratory has had at i t s disposal the prototype of a polygraphic microtelemetric system : 4 channels, 4 g r . , 3 weeks autonomy (11). At the moment, we are in a p o s i t i o n to demonstrate the processing in real-time of b i o l o g i c a l signals received by this method ( f i g u r e 5). Although adjustments may be necessary, the results are very encouraging. A miniature telemetry technique f o r e i e c t r i c a l and pharmacological stimulations w i l l suon be added, o f f e r i n g a b r i g h t future for research in psychophysioiogy and ethology.

Conclusion : We would l i k e to perfect rapid and viabie techniques of analysis f o r electrophysiologica! and metabolic a c t i v i t i e s in an animal l i v i n g as normally as possible, a necessary condition f o r the advancement of knowledge of the physiological basis of behaviour.

We wish to thank M, Rodi f o r his valuable assistance, This work was supported by D,R,M,E, (74/138), D~G,R,S,T, (75 7 0092) and C.N.R.S, grants.

139 t

l

53~427

D EOG EMG

SL

+FU+SI+SL+OS÷BP+

SI

+DS+

SI

+

PA +DP+

PA

+OS+ DP +

PA

+Op÷

Pk

171 170 550 SL 171 ~0 $ | 171 210 O$

170 590 FU 171 i 0 D$ 171 220 DP

l?l 0 S! 171 90 "-I 171 240 Pk

171 10 $L 171 150 PA 171 21)0 DP

171 20 DS 171 170 ~P [ ? t 290

1Tl 30 OP 171 180 PA

63E43Z

EOG

~

EMG

~

~

~

_

_

~_

Pk

+ PA+

PA

+ DP +

PA l?s

PA 17Z ~50 I~A 172 ~60 PA 172 ~ 0 0 P 173 50 PA 17~ ~0 yir 175 70 PA

l ? Z ~60 PA

173

20 YE

¢3~43S

HPCD

]

PA

+0P+

PA

+ O$+~+

El

+NA+

EA

174

PA 17) 440 DP 1?3 450 PA 174 50 NA 174 60 EA

~T3 J60 D$

175 570 NA

175 580 EA

RAT119 [ A ACTfVEWAKEFULNESS

PA PARAOOXICALSLEEp

Nk UNACTIVEWAKEFULNESS

YE EYES MOVEMENTS

SL: SLOWWAVES SLEEP

0S: UNCERTAINSLOWWAVESSLEEP

FU: SPINDLES

DP: UNCERTAINPARADOXICALSLEEp

~[: (NTERMEO~&TESTATE

CALIBAATION: 1$EC ;200 p~

Figure 1

DIFFERENTPHASES OF THE WAKINGSLEEPING CYCLE DETECTED BY THE OFF LiNE PROGRAM Under each of the three parts of this recording is the c o r responding l i s t i n g The f i r s t three numbers indicate the minute of recording ( m i n ) the following three give the seconds ( 5 5 0 ) On the central t r a c e note the intervention of the prohibitory logic (min 172 sec 45) which indicates the appearance of paradoxical sleep despite the electromyo gram which could have resulted in active waking Abreviations FF frontal cortex gram EMG electromyogram

HPCD dorsal hippocampus

From GOTTESMANNet a l

1973

EOG

electroocuio

140

ee*

A

~TA7$

~®

~AT

ie~ ~VE|L A~T|F

TRA~CNE

~14 $

,~O~tRNEE

~OURCE~TAGE$

EVE|L ~0~ A C T | F

PAR

$OMMEIL LENT

TRA~CH[6 6TADE OE FUSEAUA

OeO $5.S

0,0

37.6

0.0

3.6

0*2

1.3

66,?

0,0

~I,Z

0.0

~*0

0,0

0,0

6,9,1

0°0

~9~2

0.0

,Z1 ~ .

2,3

3,3

~J

DE

~1

~

Q fl | $

M~

?~

O $

$1ADE I~TERNED[A[RE

e~* $OMHE|L ~ARADOXAL

19,6

0o7

7,2

0o0

Oe4

16.?

110,7

7°5

1,5

0.0

OJO

0.0

OeO

93el

6o6

O,Z

0,0 o.o

ooo

o.o

0.0

0.0

0.0

99.9

0.0

0.0

0.1

6~!

0*3

6.8

0.0

0o!

6~*3

3~.4

;%g

O*O

6*B

0.2

~**?

0.0

0,1

70.0

29*6

2,6

ee***

ai: i

~. : L-~:-::LJ~Ik:~A ~i;i" i : 1: £--: :: .~"--:: i: ::F T "::~ : ,, T..~:::: j-TF

FiT IT : L:.i:~[~il] ::E~L L!-~I !i t .: 'i "

:: i- :l!-:] : J

:I "" u L ] - - - : . . . . . . Q ~ ! ] J ~I

-~.i

1: "--T7 1

: [ 'r-

ii

' ;7

!r--

:-i

:::il

i

i! "

i

. . . . . .

!i

w l[lllll

:

SFAUM

l:

!

i!

:

JL ' £L

k~t

F~--

IJI,J~CTIOI ::

:,

-

i~ i _.j.:'" !I :; E --

~,1~

": r--~

j

,

j

i

'

! I

I

m T

PHY~IO[D~/OLIE

~

~!U;:

:

,

:

,4~'L~

] ' i

i

I "

i

,

ii :

,

]-F-]-[

] ! I',

!

)-1 .t

IIJ l

OXOGUE : ~-! [tt

.

:

, ::[]- : ~ r

i : ::! :'qF!k ,[

! .....

~qF i

1;.IiI:i: ii !-FI:F:]g:i, ]

=

:

L '

l~ i i :. ! "t ~ - P!:i ~ [ . i]--.-i!---i..... --i ~-~-z-:~-.i..,~4~L~I-~:-L.L_L-,.~..:j IEIU~'EIIATI~I

:

SERUM ~'H?$1OLIGIILIE

Figure 2 : A/ STATES OUTPUT LISTING . Each column, indicates f o r every q u a r t e r o f an hour and every hour, the percentage o f the t o t a l time passed in each s t a t e with the percentage o f " d o u b t f u l s " at the side . The l a s t three columns coll e c t t o g e t h e r the values f o r waking as a whole, sleeping as a whole, and a l l the "doubtfuls". B/ STATES OUTPUT ON A HISTOGRAM . An example taken from the study o f a psychotropic drug . Evolution o f a c t i v e waking ( l e f t hand column) and normal waking ( r i g h t hand column) ; during the control period (top histograms), the administ e r i n g o f the drug (middle), the recuperation period (bottom) . Note, on the r i g h t , the average f o r each s t a t e f o r the duration o f recording. For a b r e v i a t i o n s , see f i g u r e I From TASSET and GOTTES~IANN (unpublished)

141

[~

N;gN

i

N:72

A

RAT

JOURNEE

214 S

ETATS

NOMBRE DtOBSERVATION5

B

DU

21

2

72

ZNDUIT$ EN , / ,

EA

NA

SL

FU

S!

PA

YE

EA

592

287

6~

29

5

O

1

2

0

NA

430

221

40

5~

5

0

0

1

0

SL

~fiO

226

4

3

60

12

20

0

0

FU

203

72

2

1

29

35

32

0

0

S!

337

167

1

0

29

20

50

0

O

PA

71

~9

?

3

20

0

1

69

0

YE

0

0

0

0

0

0

0

0

0

Figure 3 : AUTOMATICANALYSIS OF CEREBRAL REACTIVITY IN THE RAT . A/ STUDY OF EVOKEDPOTENTIALS . The upper curve shows the n~an amplitude of two successive potentials . The lower one represents the corresponding dispersion f o r th~ d i f f e r e n t components of the response . B/ ANALYSIS OF INDUCEDSTATES . In the l e f t column, the numbers indicate the number o f stimulations taken into account f o r each state f o r this analysis . On the r i g h t , written in percentages, are the states observed in the second f o l l o w i n g stimulation . Thus, 430 stimulations in normal waking were followed by normal waking in 54% of the cases, by active waking in 40 % of the cases, e t c . . . Note that this animal did not have an electrooculogram. The second column of numbers (287, 221, e t c . . . ) shows the number of samples counted f o r the evoked p o t e n t i a l analysis, the reduction being the r e s u i t of l o g i c a l prohibi t i on . For abreviations, see figure i

142

;40

Cx

744548

F

0 i~ sL 0 14 FU O I ~ SL o 17~

EMG

o 20 D 2~ e~ ~3

PR PR PR PA

0 ~

Pn

0~4~

SL

o

FU

.

St

.PAoSL,

PA

+NA.

PA

744550

!

o * ~ PR 0 4 8 P~ 0 4 9 ~A

PA

,

EA

oNA+

EA

Figure 4 : DIFFERENT PHASES OF WAKING AND SLEEPING ANALYSED BY THE REAL-TIME PROGRAM . Continuous recording with corresponding l i s t i n g on the r i g h t . The analysis has been undertaken without the organisational logic which allows the e l i m i n a t i o n of the rare aberrant states . Eye movements (MO) of paradoxical sleep are not analysed . Channel I is f i l t e r e d . For a b r e v i a t i o n s , Scale : I sec.

see figure 1

100 /~V

143

744837

CxF

D 21 22 23 2+ ~ 2+ 27 2~

MO EMG

~L SL SL SL FU ~L ~L ~C

~2 ~Z

33~L~

34 3 ~ ~L 3 ~ SL 0 3 7 SL

J

SL

.FU+

SL

.IIA.

SL

.SI.FU.

Sl

.

PA .EA.

PA

0 ~ 0 4 0 ER 041 Q + a P~

O + 6 PR 0 ~7 P. 744|~|

0 ~ e SL t~

t~ E~ E~

L

PA oSL.

Figure 5 :

PA

.SL.

E1

.IDA.

EA

.MA.

~E~

EA ,SL.EA.

REAL-TIME ANALYSIS OF A RAT RECORDED BY TELEMETRY .

Continuous r e c o r d i n g w i t h corresponding l i s t i n g on the r i g h t . As in f i g u r e 4, the processing was completed w i t h o u t the use o f l o g i c a l o r g a n i s a t i o n o f the states . Note the two s h o r t f i e l d losses in the awake r a t . Eye movements (HO) o f p a r a d o x i c a l sleep are not analysed . For a b r e v i a t i o n s , see f i g u r e 1. Scale : 1 sec.

200~V

144 1. GOTTES~,~NNCl. : Recherche sur I~ Psychophysioiogie du sommeil chez le Rat. Presses du Palais-Royal, Paris, 1967, lb6 pages. 2. GOTTESMANNCI. : Annie Psychologique, 1971, 71, 451-48U. 3. GOTTESMANNCl JUAN DE MENDOZAJ.L. LACOSTEG. LALLEMENI B. ROURIGUES L. et TASSET M. : C.R. gcad. Sciences Paris, 1971, t . 212, 301-302. 4. GOITESMANNCI., JUAN DE MENDOZAJ . L . , LACUSTE G., LALLEMENTB. RODI H., RODRIGUES L. et TASSET M. : C.R. Soc. B i o l . Paris, 1971, 165, n° 2, 373. 5. GOTTESMANNCI., JUAN DE MENDOZAJ . L . , LACOSTE G., LALLEMENTB., RODI M., RODRIGUES L. et TASSET M. : Psychophysiology, 1972, Vol. 9, n° I , 142. 6. GOTTESMANNCI., TASSET M., JUAN DE MENDOZAJ . L . , LACOSTE G., RODI M., RODRIGUES L. et ROUX R. : Symposium de Wurzburg "La Nature du Sommeil". Gustav Fischer Verlag, 19/3, 98-101. 7. GOTTES~NN C l . , LACOSTE G. JUAN DE MENDOZAJ . L . , TASSET M., RODRIGUES L.: Informatique M~dicale, Colloques IRIA, 1973, t . I , 295-309. 8. GOTTESMANNC l . , TASSET M., RALLO J . L . , JUAN DE MENDOZAJ . L . , GAUTHIER P. LACOSTE Go, RODI M., RODRIGUES L. et ROUX R. : C.R. Acad. Sciences Paris, 1972, 274, 3601-3602. 9. GOTTESMANNC I . , TASSET M., RALLO J . L . , JUAN DE MENDOZAJ . L . , GAUTHIER P., LACOSTE G., RODI M., RODRIGUES L. et ROUX R. : Psychophysiology, 1973. 10. GOTTESMANNC I . , LACOSTE G. et RODRIGUES Lo : A.P.S.S. Edinburgh, 19/5, Psychophysiology ( i n press). 1!. GOTTESMANNCI., RODI M. et RALLO J.L. : BIOCAPT Paris 1975 ( i n preparation) 12. SIEGEL S. : Non parametric s t a t i s t i c s for behavioural sciences. The Graw H i l l Co, New-York, 1956. 13. JOHNSON C. : Use of computer i n sleep research. A.P.S.S. Bruges, 1971. 14, KOHN M.~ LITCHFIELD D,~ BRANCHEYM. and BREBBIA D.R.: Electroenceph. c l i n . Neurophysiol.~ 1974, 37, 518-520. 15. BRANCHEYM., BREBBIA D.R., KOHN M. and LITCHFIELD D. : Electroenceph. c l i n . Neurophysiol., 1974, 37, 501-506.