- Email: [email protected]
Cross-correlations between discharge patterns of primary muscle spindle endings in active triceps-surae muscles of the cat
Neuroscienee Letters, 5 (1977) 63--67 © E~vier/North-Holland Scientific Publishers Ltd,
63
CROSS-CORRELATIONS BETWEEN DISCHARGE PATTERNS OF PRIMARY MUSCLE SPINDLE ENDINGS IN ACTIVE TRICEPS-SURAE MUSCLES OF THE CAT U. WINDHORST
Physiologisches Institut der Universit~t Gi~ttingen, Lehrstuhl II, GiSttingen (G.F.R.) (Received February 23rd, 1977) (Revised version received March 15th, 1977} (Accepted March 15th, 1977 )
SUMMARY
Cross-correlations were computed between simultaneously recorded spike trains from primary spindle endings lying in variably active triceps-surae muscles of decerebrate cats. The discharge patterns of these primary endings were rather frequently correlated, the correlation function usually resembling that for nearly synchronous firing. This feature is most often particularly pronounced in closely neighbouring spindles. After light cur~ization, crosscorrelations are generally diminished, but not always abolished completely, suggesting a still-existing correlating influence of gamma motoneurones. Deep curarization or deefferentation abolishes significant features in the crosscorrelograms.
Since the spinal reflex loop is a multi-channel system, it is very important for the elucidation of its function to know under which conditions and to what extent the channels are correlated; this has not yet been investigated in detail. In this report the Ia pathway from variably active triceps-surae muscles of decerebrate cats was checked for inter-channei correlations. The decerebrate preparation exhibiting alpha and gamma activity was chosen because this activity could be expected to deliver 'common input' to spindles in a muscle and thus to correlate their discharge patterns. In order to get preparations with predominant alpha activity, the cats were anaemically decerebrated by the method of Pollock and Davis [7], sometimes complemented by pre- or intracollicular midbrain section and by anterior cerebellectomy. Conventional methods were used for further operation, data acquisition and storag~ [9~. Auto- and cross-correlations were calculated on at least 3 primary endings from each experiment {tGta] of 33 pairs in 7 experiments} using a m e t h o d described in the previous communication (8 ]. Briefly, the two spike trains to be correlated were transformed into a refer-
4
A 4 - 8 6o/,0~-,
o.-
!
A .~-bA
,i 0
100
200 -80 -1;
J~ k,.
h
I['
0
o ~D
1;
B~-'C ~ 50
B'~B
u~
0
i
100
t
200
~jj
|,,
60"
I'
0
100
control
O
t
,
100
l
6o-
~
:
J
200 ms
control
....
,
.
I
-80
- ....... : i
B-'~C /!
.~
(/)
A~B
-1:
0
1;
I
|00
0
|OO
I
200 ms
O
E
C~--A 60"~
° ~
C---~A
60 l
C~--~C
.
!
0
100
.
.
control
.
0 ! , I I 200 -80 -1; 0 1: 100 0 time 1: from reference trigger
100
200 ms
Fig. 1. Auto- and cross-correlograms for three simultaneously recorded primary spindle endings (A, B and C) from the active lateral gastrocnemius muscle (drawn by an XY-plotter). Anaemic decerebration complemented by precollicular midbrain section. In the cross, correlograms the reference train is denoted by the right letter pointed at by the arrow on the right side of the ordinates. A cross-correlation for negatiw times (--1)is equivalent to the cross,correlation for positive times ( r ) w h e n the spike trains are interchanged. This fact is symbolized by the inversed direction of the arrow between 3pindle symbols on the left side of the ordinates. The three cross-correlations were computed by changing the reference spike trains in a cyclic order. The controls were computed by replacing the reference treins by artificial regular trigger trains (five superimposed runs). They yield some kind of 'confidence belt' for random fluctuations in cross-correlograms for uncorrelated trains (horizontal lines). Five hundred analysis periods of 200 msec duration; sampling rate 4000/sec; pulse width 4 msec. Mean discharge frequencies: for A 29/see; for B 38/sec; for C 26/sec.
ence trigger train and a train of rectangular pulses of preselected duration. The reference trigger train was used to trigger sampling and averaging ('analysis') periods in an averaging compu~r (Didac 800)while, in parallel, the pulse train was fed into and played back from a magnetic tape (HP 3960) in order to be delayed by 84 msec with respect to the reference trigger train. It was then fed into the averager to be sampled and averaged. This procedure yielded cross-correlations for negative times, --r(up to 84 msec) and for
65
positive times, r, from reference triggers (cf. Figs. 1 and 2). In Fig. 1, autocorrelations (symbolized by double-headed arrows between capital letters representing spike trains) and cross-correlations (symbolized by single-headed arrows) are shown for three primary endings from the lateral gastrocnemius (GL) muscle. The muscle was kept extended at 2 mm from a zero length at which it produced a low mean tension of about 20--40 g. It exhibited small amplitude tension fluctuations with occasional larger excursions. There is a more than random probability for all the endings to fire nearly synchronously, as reflected in the peaks near or about the origin of the cross-correlograms. Only the spindles A and C show a non-zero phase in their firing in that spindle C tends to fire relatively often 7--8 msec later than A. The most prominent peak is seen in the cross-correlogram between spindles B and C which were situated closely together in the proximo-lateral part of the GL muscle, whereas the third spindle (A) lay more distally in the muscle belly. It should be noted that when spindles tend to fire synchronously, the auto-correlation pattern of one ending (e.g. B*,B) reappears in the cross-correlogram with the second ending (e.g. B-+C). cur.
60, n
cur.
/.0L_
C <'-'~C
0
ffl C
_2
O_
! 100
0
o~
J 100
0
I:
1
200 rns
cur.
"0
o
7ot
Y
"~.
| 2O0
~
A*'-C
de-eft:.
A ---~C
50A<--- C
A---~ C
o
E
~r
v
m
•
i -80
I
-'~
0
1;
100
I
-80
time 1; from reference
-'t
0
1;
I 100 ms
trigger
Fig. 2. Auto- and cross-correlograms for spindles A and C of Fig. 1 after light curarization
(cur., 0.075 mg/kg) and cross-correlogram after de-efferentation (de-eft.)o Controls not reproduced. Muscle extensions of 11 mm (cur.)and 9--13 mm (de-eft.). No extrafusa| activity detectable. Three hundred analysis periods of 200 msec; sampling rate 4000/sec; pulse width 6 msec. Mean discharge frequencies: cur.: for A 25/sec, for C 10/sec; de-eft.: for A 17/see, for C 7/sec.
66
Obviously the demons~rated synchrony was at least partially caused by regional extmfusal activity since it disappeared in two correlograms (between spindles A and B, and B and C; not illustrated) after a small dose of curare (0.075 mg/kg Curarin-Asta) which just paralyzed the extrafusal muscle, but presumably left unaffected a considerable amount of intrafusal activity, However, under this condition the cross-correlogram between spindles A and C changed its appearance only slightly in that the peak shifted to r = 0 msec as illustrated in Fig. 2. The most probable explanation seems to be that these particular spindles received a common gamma input. This tentative conclusion must be confirmed by more rigorous tests, however. After de-efferentation (L6--$1), no significant features could be detected in the cross-correlograms. The demonstrated results are confirmed and complemented by those obtained in 6 other experiments. In summary, extrafusal activity may cause correlations between discharge patterns of primary endings from a muscle. The form and strength of these correlations depend on the degree of muscular activity, on muscle length and on the spatial relation of the spindles as will be documented in more detail in forthcoming papers. There are hardly ever correlations between distant spindles, particularly those in different (synergic) muscles, and in muscles with low extrafusal activity. Furthermore, it is suggested that common gamma innervation may correlate primary ending discharges which again would occur most probably in nearby spindles [1 ]. The latter effect may well be due to static gamma motoneurones which are known to elicit easily individual discharges of primary endings [2,4]. Both kinds of correlating mechanism are of the type of common input (cf. ref. 6), which is apparently confined to restricted muscle regions. In the case of extrafusal activity, this may be due to the limited territories of those motor units [ 3] whose combined actions influence the correlated spindles. V~nat may be said about the meaning of these correlations? The very fact that inter-channel correlations in the Ia pathway from an active muscle (i.e. under almost natural conditions of alpha and gamma activity) can occur at alJ~ is very important in itself. These correlations must be taken into account when the signal transmission in such a multi-channel system is considered theoretically (cf. ref. 5). Correlated channels contain a certain amount of common information generated by the source shared by them. Since this source is more or less regionally limited, the correlated Ia afferents convey information about these local events, and this is done with a higher safety factor than by single uncorrelated channels. The next question would be by which special mechanisms the CNS structures might be capable of filtering out this information from the discharge patterns of primary endings. The existence of such mechanisms is well conceivable in view of ~he high variety of combinations which are possible for the terminations c~f different Ia afferents on t~he extensive soma-dendritic system of a motoneurone, and in view of the intricate intemeuronal network of the spinal cord.
67 ACKNOWLEDGEMENTS
I am very grateful to Prof. H.-D. Henatsch and Dr. J. Meyer-Lohrnann for critical comments on the manuscript and to Mr. Howard Schultens for
checking the English text. REFERENCES 1 Adal, M.N. and Barker, D., Intramuscular branching of fusim.otor fibres, J. Physiol. (Lond.), 177 (1965) 288--299. 2 Bessou, P., Laporte, Y. and Pages, B., Frequencygrams of spindle primary endings elicited by stimulation of static and dynamic fusimotor fibres, J. Physiol. (Lond.), 196 (1968) 47--63. 3 Brandstater, M.E. and Lambert, E.H., Motor unit anatomy. In J.E. Desrnedt (Ed.), New Developments in Electromyography and Clinical Neurophysiology, Vol. I, Karger, Basel, pp. 14-~22. 4 Brown, M.C., Crowe, A. and Matthews, P.B.C., Observations on the fusirnotor fibres of the tibialis posterior muscle of the cat, J. Physiol. (Lond.), 177 (1965) 140--159. 5 Milgram, P. ~:nd Inbar, G.F., Distortion suppression in neuromuscular information transmission due to interchannel dispersion in muscle spindle firing thresholds, IEEE Trans. Biomed. Engng, BM~23 (1976) 1--15. 6 Moore, G.P., Segundo~ J.P., Perkel, D.H. and Levitan, H., Statistical signs of synaptic interactions in neurons, Biophys~ J., 10 (1970) 876--900. 7 Pollock, L.J. and Davis, L., The reflex activities of a decerebrate animal, J. comp. Neurol., 50 (1930) 377--411. 8 Windhorst, U., A simple method of cross-correlating spike trains demonstrated on primary muscle spindle ending discharges.~ Neuroscience Letters, 5 (1977) 57--61. 9 Windhorst, U., Meyer-Lohmann, J. and Schmidt, J., Correlation of the dynamic behaviour of deefferented primary muscle spindle endings with their static behaviour, Pflfigers Arch. ges. Physiol., 357 (1975) 113--122.
63
CROSS-CORRELATIONS BETWEEN DISCHARGE PATTERNS OF PRIMARY MUSCLE SPINDLE ENDINGS IN ACTIVE TRICEPS-SURAE MUSCLES OF THE CAT U. WINDHORST
Physiologisches Institut der Universit~t Gi~ttingen, Lehrstuhl II, GiSttingen (G.F.R.) (Received February 23rd, 1977) (Revised version received March 15th, 1977} (Accepted March 15th, 1977 )
SUMMARY
Cross-correlations were computed between simultaneously recorded spike trains from primary spindle endings lying in variably active triceps-surae muscles of decerebrate cats. The discharge patterns of these primary endings were rather frequently correlated, the correlation function usually resembling that for nearly synchronous firing. This feature is most often particularly pronounced in closely neighbouring spindles. After light cur~ization, crosscorrelations are generally diminished, but not always abolished completely, suggesting a still-existing correlating influence of gamma motoneurones. Deep curarization or deefferentation abolishes significant features in the crosscorrelograms.
Since the spinal reflex loop is a multi-channel system, it is very important for the elucidation of its function to know under which conditions and to what extent the channels are correlated; this has not yet been investigated in detail. In this report the Ia pathway from variably active triceps-surae muscles of decerebrate cats was checked for inter-channei correlations. The decerebrate preparation exhibiting alpha and gamma activity was chosen because this activity could be expected to deliver 'common input' to spindles in a muscle and thus to correlate their discharge patterns. In order to get preparations with predominant alpha activity, the cats were anaemically decerebrated by the method of Pollock and Davis [7], sometimes complemented by pre- or intracollicular midbrain section and by anterior cerebellectomy. Conventional methods were used for further operation, data acquisition and storag~ [9~. Auto- and cross-correlations were calculated on at least 3 primary endings from each experiment {tGta] of 33 pairs in 7 experiments} using a m e t h o d described in the previous communication (8 ]. Briefly, the two spike trains to be correlated were transformed into a refer-
4
A 4 - 8 6o/,0~-,
o.-
!
A .~-bA
,i 0
100
200 -80 -1;
J~ k,.
h
I['
0
o ~D
1;
B~-'C ~ 50
B'~B
u~
0
i
100
t
200
~jj
|,,
60"
I'
0
100
control
O
t
,
100
l
6o-
~
:
J
200 ms
control
....
,
.
I
-80
- ....... : i
B-'~C /!
.~
(/)
A~B
-1:
0
1;
I
|00
0
|OO
I
200 ms
O
E
C~--A 60"~
° ~
C---~A
60 l
C~--~C
.
!
0
100
.
.
control
.
0 ! , I I 200 -80 -1; 0 1: 100 0 time 1: from reference trigger
100
200 ms
Fig. 1. Auto- and cross-correlograms for three simultaneously recorded primary spindle endings (A, B and C) from the active lateral gastrocnemius muscle (drawn by an XY-plotter). Anaemic decerebration complemented by precollicular midbrain section. In the cross, correlograms the reference train is denoted by the right letter pointed at by the arrow on the right side of the ordinates. A cross-correlation for negatiw times (--1)is equivalent to the cross,correlation for positive times ( r ) w h e n the spike trains are interchanged. This fact is symbolized by the inversed direction of the arrow between 3pindle symbols on the left side of the ordinates. The three cross-correlations were computed by changing the reference spike trains in a cyclic order. The controls were computed by replacing the reference treins by artificial regular trigger trains (five superimposed runs). They yield some kind of 'confidence belt' for random fluctuations in cross-correlograms for uncorrelated trains (horizontal lines). Five hundred analysis periods of 200 msec duration; sampling rate 4000/sec; pulse width 4 msec. Mean discharge frequencies: for A 29/see; for B 38/sec; for C 26/sec.
ence trigger train and a train of rectangular pulses of preselected duration. The reference trigger train was used to trigger sampling and averaging ('analysis') periods in an averaging compu~r (Didac 800)while, in parallel, the pulse train was fed into and played back from a magnetic tape (HP 3960) in order to be delayed by 84 msec with respect to the reference trigger train. It was then fed into the averager to be sampled and averaged. This procedure yielded cross-correlations for negative times, --r(up to 84 msec) and for
65
positive times, r, from reference triggers (cf. Figs. 1 and 2). In Fig. 1, autocorrelations (symbolized by double-headed arrows between capital letters representing spike trains) and cross-correlations (symbolized by single-headed arrows) are shown for three primary endings from the lateral gastrocnemius (GL) muscle. The muscle was kept extended at 2 mm from a zero length at which it produced a low mean tension of about 20--40 g. It exhibited small amplitude tension fluctuations with occasional larger excursions. There is a more than random probability for all the endings to fire nearly synchronously, as reflected in the peaks near or about the origin of the cross-correlograms. Only the spindles A and C show a non-zero phase in their firing in that spindle C tends to fire relatively often 7--8 msec later than A. The most prominent peak is seen in the cross-correlogram between spindles B and C which were situated closely together in the proximo-lateral part of the GL muscle, whereas the third spindle (A) lay more distally in the muscle belly. It should be noted that when spindles tend to fire synchronously, the auto-correlation pattern of one ending (e.g. B*,B) reappears in the cross-correlogram with the second ending (e.g. B-+C). cur.
60, n
cur.
/.0L_
C <'-'~C
0
ffl C
_2
O_
! 100
0
o~
J 100
0
I:
1
200 rns
cur.
"0
o
7ot
Y
"~.
| 2O0
~
A*'-C
de-eft:.
A ---~C
50A<--- C
A---~ C
o
E
~r
v
m
•
i -80
I
-'~
0
1;
100
I
-80
time 1; from reference
-'t
0
1;
I 100 ms
trigger
Fig. 2. Auto- and cross-correlograms for spindles A and C of Fig. 1 after light curarization
(cur., 0.075 mg/kg) and cross-correlogram after de-efferentation (de-eft.)o Controls not reproduced. Muscle extensions of 11 mm (cur.)and 9--13 mm (de-eft.). No extrafusa| activity detectable. Three hundred analysis periods of 200 msec; sampling rate 4000/sec; pulse width 6 msec. Mean discharge frequencies: cur.: for A 25/sec, for C 10/sec; de-eft.: for A 17/see, for C 7/sec.
66
Obviously the demons~rated synchrony was at least partially caused by regional extmfusal activity since it disappeared in two correlograms (between spindles A and B, and B and C; not illustrated) after a small dose of curare (0.075 mg/kg Curarin-Asta) which just paralyzed the extrafusal muscle, but presumably left unaffected a considerable amount of intrafusal activity, However, under this condition the cross-correlogram between spindles A and C changed its appearance only slightly in that the peak shifted to r = 0 msec as illustrated in Fig. 2. The most probable explanation seems to be that these particular spindles received a common gamma input. This tentative conclusion must be confirmed by more rigorous tests, however. After de-efferentation (L6--$1), no significant features could be detected in the cross-correlograms. The demonstrated results are confirmed and complemented by those obtained in 6 other experiments. In summary, extrafusal activity may cause correlations between discharge patterns of primary endings from a muscle. The form and strength of these correlations depend on the degree of muscular activity, on muscle length and on the spatial relation of the spindles as will be documented in more detail in forthcoming papers. There are hardly ever correlations between distant spindles, particularly those in different (synergic) muscles, and in muscles with low extrafusal activity. Furthermore, it is suggested that common gamma innervation may correlate primary ending discharges which again would occur most probably in nearby spindles [1 ]. The latter effect may well be due to static gamma motoneurones which are known to elicit easily individual discharges of primary endings [2,4]. Both kinds of correlating mechanism are of the type of common input (cf. ref. 6), which is apparently confined to restricted muscle regions. In the case of extrafusal activity, this may be due to the limited territories of those motor units [ 3] whose combined actions influence the correlated spindles. V~nat may be said about the meaning of these correlations? The very fact that inter-channel correlations in the Ia pathway from an active muscle (i.e. under almost natural conditions of alpha and gamma activity) can occur at alJ~ is very important in itself. These correlations must be taken into account when the signal transmission in such a multi-channel system is considered theoretically (cf. ref. 5). Correlated channels contain a certain amount of common information generated by the source shared by them. Since this source is more or less regionally limited, the correlated Ia afferents convey information about these local events, and this is done with a higher safety factor than by single uncorrelated channels. The next question would be by which special mechanisms the CNS structures might be capable of filtering out this information from the discharge patterns of primary endings. The existence of such mechanisms is well conceivable in view of ~he high variety of combinations which are possible for the terminations c~f different Ia afferents on t~he extensive soma-dendritic system of a motoneurone, and in view of the intricate intemeuronal network of the spinal cord.
67 ACKNOWLEDGEMENTS
I am very grateful to Prof. H.-D. Henatsch and Dr. J. Meyer-Lohrnann for critical comments on the manuscript and to Mr. Howard Schultens for
checking the English text. REFERENCES 1 Adal, M.N. and Barker, D., Intramuscular branching of fusim.otor fibres, J. Physiol. (Lond.), 177 (1965) 288--299. 2 Bessou, P., Laporte, Y. and Pages, B., Frequencygrams of spindle primary endings elicited by stimulation of static and dynamic fusimotor fibres, J. Physiol. (Lond.), 196 (1968) 47--63. 3 Brandstater, M.E. and Lambert, E.H., Motor unit anatomy. In J.E. Desrnedt (Ed.), New Developments in Electromyography and Clinical Neurophysiology, Vol. I, Karger, Basel, pp. 14-~22. 4 Brown, M.C., Crowe, A. and Matthews, P.B.C., Observations on the fusirnotor fibres of the tibialis posterior muscle of the cat, J. Physiol. (Lond.), 177 (1965) 140--159. 5 Milgram, P. ~:nd Inbar, G.F., Distortion suppression in neuromuscular information transmission due to interchannel dispersion in muscle spindle firing thresholds, IEEE Trans. Biomed. Engng, BM~23 (1976) 1--15. 6 Moore, G.P., Segundo~ J.P., Perkel, D.H. and Levitan, H., Statistical signs of synaptic interactions in neurons, Biophys~ J., 10 (1970) 876--900. 7 Pollock, L.J. and Davis, L., The reflex activities of a decerebrate animal, J. comp. Neurol., 50 (1930) 377--411. 8 Windhorst, U., A simple method of cross-correlating spike trains demonstrated on primary muscle spindle ending discharges.~ Neuroscience Letters, 5 (1977) 57--61. 9 Windhorst, U., Meyer-Lohmann, J. and Schmidt, J., Correlation of the dynamic behaviour of deefferented primary muscle spindle endings with their static behaviour, Pflfigers Arch. ges. Physiol., 357 (1975) 113--122.