Brain Research, 559 (1991) 241-248 (~) 1991 Elsevier Science Publishers B.V. All rights reserved. 0006-8993/91/$03.50 ADONIS 0006899391169815

241

BRES 16981

Favored patterns in spontaneous spike trains Yun-Hui Ku and Xian-Qun Wang Department of Physiology, Beijing Medical University, Beijing 100083 (People's Republic of China) (Accepted 30 April 1991)

Key words: Favored pattern; Spike train; Nucleus raphe magnus; Locus ceruleus

By using the modified detection method, favored patterns can be detected in a total of 44 spontaneous spike trains. Among these the 'periodical burst' discharge of one sympathetic preganglionic neuron and the 'fast-slow' alternative discharge of some hypothalamic neurons have visible characteristics, hence we use them to test the reliability of our method by comparing the detected patterns with the non-sequential interval histograms and oseillograms of the spike trains. The comparisons show that our method is reliable. The spike trains of nucleus raphe magnus (NRM) and the locus coeruleus (LC) have no visible characteristics; from these the following results have been observed: (1) all spike trains have one or more favored patterns; (2) some spike trains from neurons in the same nucleus have common fragments of favored patterns; (3) the favored patterns in spike trains recorded from different nuclei are different from each other; (4) some favored patterns in spike trains of the NRM neurons remain unchanged from beginning to end in 35-rain records and their repetitions are relatively stable; and (5) mieroinjeetion of normal saline or normal serum into the LC has no significant influence on the occurrence of favored patterns in 35-rain records of spike trains of the LC neurons. The above results indicate that the favored patterns in spike trains are objective and regular phenomena with relative stability. It seems likely that favored pattern may be used (as an index of the neuronal activity) in combination with the mieroinjeetion technique, etc., for various studies including studies on neural coding. INTRODUCTION W e have o b s e r v e d that some r e c o r d e d spike trains have regular and characteristic interval p a t t e r n s 11, which m a y r e p r e s e n t a neural code, i.e. interval patterns of spike trains m a y contain information. In o r d e r to pick up m o r e information from the spike trains, we present a modified m e t h o d for the detection o f favored patterns (FP) in spike trains after D a y h o f f and G e r s t e i n 4'5, in which we c a r d e d out studies with simulated spike trains (containing k n o w n i n t e r p o l a t e d patterns) and real spike trains. In the course of applying this m e t h o d to the real spike trains, s o m e modifications and supplements were m a d e to i m p r o v e the s p e e d o f analysis and the p a t t e r n detection rate, and to m e e t the requirements o f practice. W i t h the p u r p o s e of making sure the prospect of this modified m e t h o d and the significance of F P in the study o f neural coding, the present study was u n d e r t a k e n to test further the reliability, the availability of the m e t h o d and the stability of favored patterns d e t e c t e d in spike trains. (1) We selected the 'periodical burst' discharge of one sympathetic preganglionic n e u r o n in which patterns were visible, and the ' f a s t - s l o w ' alternative discharge o f some hypothalamic neurons with visible characteristics, for testing the reliability and availability of the modified m e t h o d b y comparing the d e t e c t e d favored patterns with

the non-sequential histograms and oscillograms of these spike trains. (2) We also a p p l i e d the m e t h o d to the spike trains of the nucleus r a p h e magnus ( N R M ) and the locus coeruleus (LC) with no visible characteristics in ord e r to see if this m e t h o d is also applicable to them. (3) We further examined the occurrences o f FPs in long (35 rain) records of spike trains and the influences of microinjecting saline o r normal serum into the same nucleus of recording on FPs, i.e. to test the stability of FPs in long spike trains. MATERIALS AND METHODS

Modified FP detection method The detection method of FP based on that used by Dayhoff and Gerstein includes the quantized Monte Carlo method and the template method (for details see ref. 4). In applying the method to the real spike trains recorded previously from neurons in the spinal cord and 6 brain areas, the following modifications and supplements were made. (1) Before using the quantized Monte Carlo method, the favored interval lengths were selected by analyzing the interval histogram to speed up following detections. (2) In view of the fact that some neural spike train data exhibit 'bursts' (having very short intervals), and some alternate between a fast and slow firing mode, we treated the fast firing (including 'bursts') and slow firing mode separately: the range for the fast firing was from several tens of microseconds to several milliseconds, that for the low firing was from several tens of milliseconds to several seconds. (3) The template method: this method is based on a 'sliding'

Correspondence: ¥.-H. Ku, Department of Physiology, Beijing Medical University, Beijing 100083, People's Republic of China.

242

(a)

..... .

.

.

III !l .

.

.

.

.

.

.

.

lllql . . . . . .

.

.

lllll ........

.

-

JL_

-lilt I ..........

[]11] - -

III11

1111

~ --"-------- ~ I111[. . . . . . ' - - - - - r - , - : ; - - - I l l ] - - - ~ - - - -

1 sec (b)

I

I

80-

t.

~4-

° o

4c-

"6

32-

o

115 E

hlill I I i III

D

Z

0 =

•

0

I

i

10

Quantized

II1 I

i

II~lh l

20

i

I

30

interval

I i

I

40

50

length

Fig. 1. a: the photograph of a spike train from the rat sympathetic preganglionic neuron, b: the interval histogram of a spike train from the sympathetic preganglionic neuron. Bin = 35 ms.

comparison of the chosen template with the entire spike train, which can determine both the locations of patterns matched with the template pattern in the spike train and the number of matches; a statistical test (the shuffling test) at the 1% significance level was used. The number of pattern repetitions is equal to the number of matches in the unshuf~ed spike train minus the maximum number in 99 shuffles. The template method which we used was slightly different from

(a)

L L_I.J.L,LhlllLd,i ..... I..... 1,

[1. I L, IL

111 ll]rl]lIrp rII"ll

1 sec t

I

that used by Dayhoff and Gerstein. Dayhoff and Gerstein used a weighting function (based on the template) fpr the detection of imperfect matches, and chose the first spike in the template to initiate the comparison. Our method allows intervals of the template to be matched in sequence with the corresponding intervals of the spike train respectively, using a certain tolerance of variation (0,10.2 of the interval length). If the successive interval lengths of the matched fragment of a spike train are all within the tolerance of the corresponding interval lengths of the template, the fragment of the spike train is said to 'match' the template. In the present study it means that, if a section of the spike train approximately matches the template pattern (i.e. the former may contain one extra or one missing spike), it can still be considered a match.

Source of spike trains All spike trains used were already recorded extracellularly in our previous electrophysiological study s'9'11 in urethane- (for the sympathetic preganglionic neuron) or chloral hydrate-anesthetized (for neurons in brain areas), curare-immobilized and artificially ventilated rats. (For details on the surgical operation, recording techniques and location of recording sites please refer to the corresponding papersS'9'll.)

c~ 280; I D

224

92.0 •

16s

e- 736 .

0

552

112

g

o 368

E "1 Z

0

"m

•

0

~ .::

10

Ouantized

l;

dUal=ill =

I =

20 interval

a

I

I

184

I

a

30

m

40

t

=

50

length

Fig. 2. a: the photograph of a spike train friom the rat hypothalamic neuron, b: the interval histogram of the spike train from the hypothalamic neuron no. 3. Bin = 35 ms.

E D Z

0

I, d. i

0

i

I

10

Quantized

i

i

I

20 interval

I

I

30

I

40

I

I

5O

length

Fig. 3. The interval histogram of a spike train from a NRM neuron (RM 20-1). Bin = 35 ms.

243 TABLE I

Favored patterns in one spike train characterized by periodical burst discharge recorded from a sympathetic preganglionic neuron * In detection, according to different condition, it is necessary to use different quantization bin width for quantization analysis; but in order to have a good comparision, the favored patterns detected were re-quantized at the same quantization resolution. ** The values of long and short interval lengths in this row are the means of both, measured respectively from photograph of spike train in Fig. la.

Quantization bin width* (ms)

Favored pattern or favored internal length

Methods of detection

Real interval lengths (ms) Measurement made from photograph of spike train**

50

1310

Quantized interval length 1

37

35

Probability distribution of intervals Intervals Probability (%) Favored pattern detection method

40 70 110 1300 1340 1370 1400 222010 5 4 7 5

1 2 3 37 38 39 40 2220105 4 7 5

35

43, 38, 1438 59, 56, 1357 25, 41, 1393 1400, 47, 49 38,40,76 67,43,38 41, 38, 38 56, 56, 1402, 47, 49 38, 36, 1341, 65, 52 50, 36, 1375, 45, 40

1, 1, 41 2, 2, 39 1, 1, 39 40, 1, 1 1,1,2 2,1,1

35

1, 1, 1 2, 2, 40, 1, 1 1, 1, 38, 2, 1 1, 1, 39, 1, 1

TABLE II

Favored patterns in spike trains characterized by fast-slow alternative discharge from the rat hypothalamic neurons Only part of the FPs are displayed in the table, the rest of FPs are not displayed due to the limitation of the table size. - , represents a long interval. HP, nucleus hypothalamicos posterior; RE, nucleus reuniens; HPE, nucleus periventricularis hypothalami; AR, nucleus areuatus.

No.

Average firing rate (spikes~s)

No.

Favored pattern Sequence of interval length (ms)

Sequence of quantized interval length

0.072, 0.072, 0.072, 0.072, 0.054, - , 0.072, 0.072, 0.072 -, 0.054, 0.036, 0.036, -, 0.054, 0.054, -, 0.036, 0.036 0.036, -, 0.054, none 0.036, 0.036, -, 0.036, 0.054 0.036, 0.054, -, 0.072, 0.036, - , 0.036, 0.036 none 0.054, - , -, 0.036, -

1, 1, 1, -, -, 1, 1, -, 1, -, 1, -, none 1, 1, 1, 1, 1, - , none 1, - ,

none 164, 70, 230, 106 286, 130, 146, 304 222, 73, 49, 217, 204 109, 73, 205, 63 53, 27, 47, 25 47, 50, 55, 46, 51, 41 493, 192, 148

none 5, 2, 6, 3 8, 4, 4, 8 6, 2, 1, 6, 6 3, 2, 6, 2 1, 1, 1, 1 1, 1, 1, 1, 1, 1 14, 5, 4

Quantization bin width (ms)

Number of FPs in one spike train

0.072

3

Fast discharge 1 (RE)

3.0

1 (RE) 2 (RE)

2 (RE)

1.4

3 (AR)

3 (AR)

3.5

4 (AR) 5 (HPE)

4 (AR)

6.0

5 (HPE)

1.1

6 (HP) 7 (HP) 8 (HP)

1, 1, 1, 1, 1, 1,

1, 1, 1 1 -

2 3 0 3

- , 1, 1 - , 1, 1, 1

4 0 1

-, 1, -

Slow discharge 6 (HP)

5.2

1 (RE)

2 (RE) 7 (UP)

13.2

3 (AR) 4 (AR)

1.6

6 (HP) 7 (HP)

5 (HPE) 8 (HP)

8 (HP)

35

0 3 5 1 3 2 1 2

244 The sympathetic preganglionic neurons at the T 2 level of the spinal cord were identified by stimulating the cervical sympathetic truck and recording the antidromic action potential (according to 3 criteria: i.e. fixed latency, ability to follow high frequency stimulation and collision with spontaneously occurring action potentials) ~1. Among them a few are characterized by a 'periodical burst' discharge (see Fig. la); we chose one such spike train for detecting FP. A part of the hypothalamic neurons studied previously present 'fast-slow' alternative discharges (see Fig. 2a) (unpublished data). Their spike trains are also suitable for examining the reliability and availability of the FP detection method. These neurons are distributed in the nucleus periventricularis hypothalami (HPE), the nucleus hypothalamicus posterior (HP), the nucleus reuniens (RE) and the nucleus arcuatus (AR). The nucleus raphe magnus (NRM) is located at the center of the rostral ventral medulla, while the LC is in the pontine and close to the fourth ventricle. The favored patterns were detected every 5 min from long spike trains (35 min) of NRM and LC. For spike trains of LC neurons, immediately after the first 5 min, normal saline or normal serum (0.04/zl) was injected over a period of 5 rain via a micropipette, 160-180/zm apart from the recording microelectrode. Sampling The spike trains recorded were fed to a window discriminator (model DSC-3), after the discharge of a single neuron was isolated and transformed into rectangular impulse trains; the output of DSC-3 was led to a microcomputer (Super PC/AT), by which the interval histograms were made and FPs were detected. The sam-

piing point in the computer is 18/zs. Each sampling spike train includes 3-5 min of activity. RESULTS Favored patterns in spike trains o f neurons in different nuclei T h e FPs in the 'periodical b u r s t ' discharge of the sympathetic p r e g a n g l i o n i c n e u r o n c a n be recognized visibly f r o m the oscillogram (Fig. l a ) . I n o t h e r words, the 'periodical b u r s t ' firings are just the FPs. T h e i n t e r v a l hist o g r a m of such a spike train reveals 2 peaks (Fig. l b ) . T h e high o n e is located in the r a n g e of short intervals, a n d the low o n e in the r a n g e of long intervals, suggesting the FPs consist of short a n d l o n g intervals, m a i n l y the short ones. Table I shows that the p r e d o m i n a n t interval lengths of the FPs d e t e c t e d b y the modified m e t h o d are in the range of 2 5 - 7 6 ms a n d 1341-1438 ms; this is e x t r e m e l y consistent with the 2 p e a k s in the interval hist o g r a m ( 4 0 - 7 0 ms a n d 1300-1400 ms) a n d the short ( a b o u t 50 ms) a n d long ( a b o u t 1310 ms) interval lengths m e a s u r e d from the oscillogram. T h e 'fast a n d slow' alternative discharge of the hypo-

TABLE III Favored patterns in spike trains from the nucleus raphe magnus (RM) neurons Only part of the FPs are displayed in the table, the rest are not displayed because of the limitation of the table size. The spike trains from neurons in the same animal were recorded 30 min apart. ' ', ' ' and ' ~ ' mark the different common favored fragments respectively. Value in parentheses represents the number (n) of pattern repetitions per 1080 spikes, n = c - m, where c is the number in the unshuffled spike train, while m is the maximum number in 99 shuffles. RM exp. no./ neuron no.

Firing rate (spikes/s)

1 RM20-1 2 RM21-1 3 -2 4 -3 5 -4 6 RM22-1 7 -2 8 -3 9 RM46-1 10 -2 11 -3 12 RM47-1 13 -2 14 -3 15 RM48-1 16 -4

9.7 5.1 6.2 9.3 5.0 0.8 6.4 12.0 3.1 0.8 3.4 1.0 0.5 2.2 16.1 9.2

17 RM49-1 18 -2 19 RM50-1 20 RM64-1 21 -2 22 -3 23 -4

12.3 8.0 5.0 5.8 1.9 4.7 2.5

Favored patterns with common fragment Sequenceof interval length (ms)

Sequence of quantized interval length

68, 68, 70, 68, 67, 67, 67, 67, 67 92, 65, 86, 83, 121 92, 87, 86, 117, 90, 124 131, 94, 94, 63, 81, 67 87, 83, 67, 94, 99 175, 192, 349, 362, 223 67, 65, 67, 67 88, 89, 90, 90, 91, 89, 87, 88, 86 164, 266, 182, 164 306, 248, 256, 293 95, 94, 97, 87, 87, 83 225, 103, 95, 74, 70, 59 52, 43, 49, 54 166, 180, 599, 419, 225 74, 103, 77, 68, 76, 74 63, 72, 72, 72, 65, 38 113, 113, 113, 108, 108 92, 85, 86, 85, 86 161, 212, 167, 218, 152, 151,220 192, 180, 184, 178, 178 158, 113, 110, 119, 112 422, 602, 773 122, 108, 110, 110, 110, 103 121, 115, 121, 115, 122

2, 2, 2, 3, 3, 4, 3, 3,

2_, 2_, 2, 2, 2 (47) 2, 2, 2, 3 (21) 2, 2, 3, 3, 3 (5) 2_, 2, 2(3) 2-, 2, 2, 3, 3 (17) 5 , 5 , 10, 10, 6 (17) 2, 2, 2-, 2(52) 3, 3, 3, 3, 3, 3, 3, 3, 3 (33) 5, 8, 5 , 5 (32) 9, 7, 7, 8 (12) 3, 3, 3, 2-, 2, 2(15) 6, 3, 3, 2-, 2, 2 (6) 2, 1, 1, 2 (41) 5 , 5 , 17, 12, 6 (14) 2, 3, 2, 2_, 2, 2(72) 2, 2, 2, 2-, 2, 1 (2) 3, 3, 3, 3, 3 (27) 3, 2, 2, 2, 2 (56) 5, 6, 5, 6, 4, 4, 6 (13) 5, 5, 5 , 5 , 5 (62) 4, 3 3~.~._.~ (7) 12, 17, 22 (36) 3, 3, 3, 3, 3, 3 (27) 3, 3, 3, 3, 3 (29)

Quantization om width (ms)

Number of FPs in one spike train

35

5 6 3 4 4 4 5 2 3 2 2 l 1 3 3 7

245 TABLE IV Favored patterns in spike trains from the locus coeruleus (LC) neurons Only part of the FPs are displayed in the table, the rest of the FPs are not displayed due to the limitation of the table size. The spike trains from neurons in the same animal were recorded 30 rain apart. ' ', ' ', ' ', and ' ~ ' mark the different common favored fragments respectively. Value in parentheses represents the number of pattern repetitions per 1000 spikes.

Group

I

II

LC exp. no. -neuron no.

Firing rate (spikes~s)

1 LC108-1 2 -2 3 LCll0-1

3.6 1.5 2.8

4 -2 5 LC137-1

1.0 3.3

1 LC100-1 2 -2 3 LC137-3 4 -4 5 LC138-1

9.2 4.2 11.6 14.5 22.6

6 7

29.2 34.8

-2 -3

Favored patterns with common fragment

Number of FPs in one spike train

(170) (90) (34) (19) (23) (35)

90

11 12 18

3, 2, 3, 2_ (27) 2.,_2,_2 (98) 5, 5, 5, 5, 5, 5 (75) 5, 5, 5, 5, 5 (49) 3, 2_, 3, 2(85) 3, 2.,_2,__2 (82) 2,2,2,2,2 (112) 2 , ~ (83)

18

Sequence of interval length (ms)

Sequence of quantized interval length

220, 215, 272, 223, 206, 264,

2, 2_, 2_ 34, 6 2, 2, 2_, 2 3 , 4 , 5, 2 2,1,1 2, 2, 1, 1, 3

236 380, 532 234, 224 387, 480, 178 142, 77, 122 139, 135, 122, 117, 248

60, 43, 65, 43 47, 43, 41 88, 88, 85, 86, 62, 44, 62, 45 61, 44, 44, 44 32, 31, 32, 40, 31, 31, 31, 31,

86, 90, 90, 88 88, 86, 90 32 32

thalamic neurons scattering in 4 nuclei can also be recognized from the oscillogram (Fig. 2a), but not very clearly, hence to measure its FPs directly from the oscillogram is difficult. But its interval histogram also reveals 2 peaks (Fig. 2b). Correspondingly, 2 groups of the FPs (fast and slow) were detected (Table II). In addition, Table II shows that 5 out of 8 spike trains have fast FPs and the common fragment (the latter was quantized as '1,1'), 7 out of 8 spike trains have slow FPs. In general the slow FPs are variable (only one pair of spike trains have the common fragment). The spike trains of NRM and LC have no visible characteristics, hence we have no prior knowledge. However, they all have FPs which can be detected by the modified detection method. Table III shows that the FPs can be found in all 23

(a)

~

Quantization bin width

5 4 6 6 10 7 2 13 9

spike trains (firing rate 0.5-16.1 spikes/s) of NRM neurons from 9 rats, and the number of different FPs in each spike train ranges from 1 to 7. Among the 23 spike trains, 10 spike trains (from 7 rats) have a 3-interval quantized common FP fragment (CFF) (2,2,2), 5 spike trains (from 3 rats) have a 4-interval CFF (3,3,3,3) and 4 spike trains (from 4 rats) have a 2-interval quantized CFF (5,5) (shown in Tables III and V). Twelve spike trains of LC neurons analyzed can be divided into 2 groups according to their firing rate: i.e. lower (group I: n = 5, 1.5-3.6 spikes/s) and higher (group II; n = 7, 4.2-34.8 spikes/s) firing rate groups. Correspondingly, there are 2 types of quantized interval histograms (Fig. 4a,b). Their peaks are located in the range of short and long intervals, respectively, suggesting that these 2 groups of spike trains might belong to 2

(13) 50

610

40

488

so

366 -

0

~

aO

,'~ E z

10 0

i llt |

0

•

•

10

244122

o- Ih,,

|

•

20

•

•

30

•

•

I

40

I

50

Quantized

0 interval

=

10

I

I

20

I

I

30

I

/

40

I

]

50

length

Fig. 4. a: the interval histogram of a spike train from an LC neuron (group I, LC 108-1); bin = 20 ms. b: the interval histogram of a spike train from an LC neuron (group II, LC 138-2); bin = 20 ms.

246 TABLE V The common favored fragments (CFF) in spike trains from the locus coeruleus (LC) and nucleus raphe magnus (RM) neurons

Values in parentheses represent the number of pattern repetitions per 1000 spikes. Nucleus

Group

LC

I II

Number of neurons

RM

Quantization bin width (ms)

Quantized common favored fragment

Neurons with common favored fragment

5

90

7

18

2, 3, 2, 3, 2, 5, 2,

LC108-1 LC108-2 LCl10-2 LC109-1 LC109-2 LC137-3 LC138-2

(170), (90), (23), (27), (98), (75), (112),

23

35

2, 2, 2

RM20-1 RM21-1 -3 -4 RM22-2 RM22-1 RM46-1 RM22-3 RM48-4

(47), (21), (3), (17), (52), (17), (32), (33), (27),

2, 4 1, 2, 2, 5, 2,

2 1 3, 2 2 5, 5, 5 2, 2, 2

5, 5 3, 3, 3, 3

kinds of neurons or repesent 2 kinds of functional status. FPs can be found in all 12 spike trains, but the number of different FPs in each spike train ranges from 2 to 18. In both groups of spike trains, there are pairs of spike trains which share the same FP fragment (the common fragments ranging from 3 to 5 intervals). The above results indicate that the method used is reliable and appropriate for detecting various spike trains o f different neurons. Variations in the CFF repetitions (Fig. 5, Table VI) Four spike trains with a higher number of CFF (quantized CFF: 2,2,2) repetitions were chosen for measurement of variations. Table VI shows that the interval lengths of the same CFF are very d o s e , and the largest CV (coefficient of variation) of interval length is 12.5% which is very close to that of 10 CFF templates (13.0%,

LCll0-1 (34) LCll0-1 (19) LC137-1 (35) LC138-1 (85) LC138-1 (82) LC137-4 (49) LC138-3 (83) RM46-3 RM47-1 RM48-1 -4 RM49-1 RM47-3 RM50-1 RM64-1 -3 -4

(15) (6) (72) (2) (56) (14) (62) (33) (27) (29)

Number of neurons with CFF

2 2 2 2 2 2 2 10

4 5

Fig. 5). But there is a high individual variation in the number of CFF repetitions as shown by the range of it (from 2 to 72 per 1000 spikes) in 10 different spike trains (Fig. 5). The stability o f FPs in long spike trains The same FPs could be detected every 5 rain from 35rain recordings of spontaneous discharge of 4 N R M neurons and remained quite stable (Fig. 6a), indicating that FPs are not stochastic p h e n o m e n a presenting in a short period, but objective and regular phenomena. After recording the first 5-rain discharge of an L C neuron, normal saline or normal serum was injected into the same nucleus to test its influence on FPs in long spike trains of LC. The same FPs could also be found every 5 min from 35-rain L C spike trains and they remained stable (Fig. 6b), thus not only further demonstrating the stability of FPs but also showing that normal

TABLE VI The variations of the CFF (common favored fragment) repetitions in 4 spike trains

* The CFF repetitions containing one extra or one mi~ing spike are excluded. The largest CV (coefficient of variation) of interval length in the table is 12.5%. Interval lengths o f pattern repetitions

RM exp. no./ neuron no.

Firing rate (spikes/s)

Template CFF (ms)

CFF repetitions* (timeMnumber of spikes)

7¢ +- S.E.M. (ms)

CV (%)

1 2 3 4

9.7 6.4 16.1 12.3

68, 65, 68, 85,

19/1494 11/965 31/996 9/1039

68.0 63.9 70.3 83.5

8.3, 8.5, 8.9 10.4, 12.5, 11.9 11.1, 10.6, 9.6 7.6, 8.5, 11.7

RM20-1 RM22-2 RM48-1 RM49-1

67, 67, 76, 86,

67 67 74 85

- 1.3, 67.1 --+ 1.3, 68.5 -+ 1.4

-+ 2.0, 66.1 --+ 2.5, 69.5 + 2.5 -+ 1.4, 78.7 +-- 1.5, 75.3 +-- 1.3 +- 2.1, 84.7 +- 2.4, 87.4 +- 3.4

247 Quantization Quontized

binwidth

: 3~ ms

CFF:

2

1RM20-~

I

2 RM 21-1

I I

7RM

47-1

8RM4s-1 lO RM 49-1

I I

I

e7

CV

47

63

I

el

I

~7

I

3

I

6s

I

67

I

17

65

I

I

67

83

I

I

67

67

52

I

e3

74

I

70

I

e6

I

re

I

74

I

72

I

72

I

72

I

6~

I

59

I

8e

73.4"1"3.0 .

7 7 . 5 -+ 2 . 6 .

13.0"1,

10.4 "1,

21

I

I

n =10 M e a n t S.E.

I

I

I

9 RM48-4

67

(Times / 1000 spikes )

2

86

67 I

,

I

I

I

I

2

65

67

5RM22-2

eRM 4 e - 3

e8

I

3RM21-3 4RM21- 4

CFF repetitions

16

I

6

2

6~

I

I

I

5 0 ms

~e

72.4::!: 2.8 12.0 "1,

Fig. 5. The common favored fragments (CFFs) of 10 RaM spike trains and the number of their repetitions. * The CFF repetitions containing one extra or one missing spike are included. The number of CFF repetitions = c - m, where c is the number in the unshuf~ed spike train, while m is the maximum number in 99 shuffles.

saline or normal serum has no significant influence on FPs in spike trains. DISCUSSION

The first purpose of the present study was to test the reliability and availability of the modified detection method. By comparing the detected FPs in the 'periodical burst' discharge of the sympathetic neuron and the 'fast-slow' alternative discharge of the hypothalamic neurons with the oscillograms and the non-sequential histograms of these spike trains, it has been proved that the modified method is reliable. These results together with those from the spike trains (having no visible characteristics) of NRM and LC show that the method used is applicable to various spike trains of neurons. Thus, the first purpose has been fulfilled. Our second purpose was to examine the stability of FPs in long spike trains. The results demonstrate that FPs can be found from the beginning to the end in long spike trains and that the number of pattern repetitions is relatively stable, not influenced by locally injecting normal saline or normal serum into the same nucleus of recording, thus proving that the FPs are objective regular phenomena and may represent neural codes, and suggesting that FPs can be used (as an index of the neuronal activity) in combination with the microinjection technique, etc., for various studies, particularly for studying neural coding.

In addition to the above, the following results have been observed. (1) All spike trains have FPs, and one spike train usually has several different patterns. We surmise that it may be due to neurons in various nuclei being subjected to the influences from other neurons inside and outside the nucleus and making different responses at the same time. (2) The FPS in spike trains recorded from different

,o0[,o, ..M OL T

(n=4)

E

o 200 (3. 150 F ~oo

rain

b)

LC

L..i

o,,no

o_ 5c I

o

0.04 ,U] (i.LC)

~

~'o ~

I

2'0 ~'5 30

rain

Fig. 6. a: showing the relative stability of the pattern repetitions in the long spike trains of the NRM neurons, b: normal saline- or serum-injection into locus coeruleus (i. LC) has no siotmificant influence on the stability of the pattern repetitions in the long spike trains of the LC neurons. Vertical bars represent S.E.M. Difference from control (time 0, t-test): all are P > 0.1.

248 nuclei are different from each other, but some spike trains r e c o r d e d from the same nucleus share the same FP fragment. The presence of c o m m o n fragments further proves that FPs are non-stochastic p h e n o m e n a and suggests that the spike trains sharing the same F P fragment might belong to the same kind of neurons or represent the same functional status. (3) Interestingly, 10 of 23 N R M spike trains have the c o m m o n favored fragment, although the N R M consists of diverse neurons releasing various transmitters 2. O n the contrary, the L C of the rat is mainly c o m p o s e d of noradrenergic neurons 3, but pairs of L C spike trains share the same F P fragment. P r o b a b l y the fact that N R M is the important relay nucleus of the descending

pain inhibition system 6'1° and L C receives diverse inputs and projects diversely to m a n y brain areas can be used to explain the causes of these paradoxical results, but their exact mechanisms are yet to be determined. Finally, it should be mentioned that in the present study only spontaneous spike trains were analyzed; the e v o k e d spike trains are being studied and the results will be r e p o r t e d later.

REFERENCES

7 Foote, S.L., Bloom, F.E. and Aston-Jones, G., Nucleus locus coeruleus: new evidence of anatomical and physiological specificity, Physiol. Rev., 63 (1983) 844-914. 8 Gao Yuang-Sheng and Ku Yun-Hui, Mechanism underlying the inhibitory effect of rat nucleus arcuatus hypothalami on unit discharge of locus coeruleus with reference to elcctroacupuncture, Acta Physiol. Sin., 35 (1983) 163-171. 9 Gao Yuang-Sheng and Ku Yun-Hui, Mechanism underlying the excitatory effect of nucleus arcuatus hypothalami on PAG-NRM system and its significance in electroacupuncture in rats, Acta Physiol. Sin., 35 (1983) 409-415. 10 Mayer, D.J., Endogenous analgesia system: neural and behavioral mechanism. In J.J. Bonica, J.G. Liebeskind and D.G. Albe-Fessard (Eds.), Proceedings of the Second World Congress on Pain, Raven Press, New York, 1979, pp. 385-410. 11 Zhou Zheng-Feng and Ku Yun-Hui, Effect of electrical stimulation of rostral ventrolateral medulla on unit discharges of sympathetic preganglionic neurons in the rat, Acta Physiol. Sin., 39 (1987) 116-122.

1 Amaral, D.G. and Sinnamon, H.M., The locus coeruleus: neurobiology of a central noradrenergic nucleus, Prog. Neurobiol., 9 (1977) 147-196. 2 Bowker, R.M., Westlund, K.N., Sullivan, M.C., Wilber, J.E and Coulter, J.D., Descending serotonergic, peptidergic and cholinergie pathways from the raphe nuclei: a multiple transmitter complex, Brain Research, 288 (1983) 33-48. 3 Dahlstrom, A. and Fuxe, K., Evidence for the existence of monoamine containing neurons in the central nervous system. I. Demonstration of monoamine in the cell bodies of brainstem neurons, Acta Physiol. Scand., Suppl. 232 (1964) 1-55. 4 Dayhoff, J.E. and Gerstein, G.L., Favored patterns in spike trains. I. Detection, J. Neurophysiol., 49 (1983) 1334-1348. 5 Dayhoff, J.E. and Gerstein, G.L., Favored patterns in spike trains. II. Application, J. Neurophysiol., 49 (1983) 1349-1363. 6 Fields, H.L., Basbaum, A.I., Clanton, C.H. and Anderson, S.D., Nucleus raphe magnus inhibition of spinal cord dorsal horn neurons, Brain Research, 126 (1977) 441-453.

Acknowledgements. This work was supported by a Grant from the State Education Commission. We thank Ms. Hong Liang and Dr. Xiao-Ying Chen (Institute of Electronics, Academia Siniea) for their helpful advice in mathematical modeling and computer programming.