Neuroscience Vol. 48, No. 4, pp. 915-924, 1992 Printed in Great Britain

BACKGROUND

0306-4522/92 $5.00 + 0.00

Pergamon Press Ltd 0 1992IBRO

FIRING ACTIVITY IN GUINEA-PIG NEOCORTEX IN VITRO S. V. KARNUP

Institute of Cell Biophysics, The Russian Academy of Sciences, Pushchino, 142292, Russia Abstract-The background firing activity was recorded extracellularly in experiments on guinea-pig neocortical slices maintained in vitro. The following types of background firing activity were revealed: (i) high regular single spikes (48%), (ii) irregular single spikes (lS%), (iii) bursts (7%), (iv) groups (7%), (v) mixed activity where single spikes alternated with bursts or groups (28%). The specific interspike interval distribution and the specific shape of autocorrelogram corresponded to each of these background firing activity types. Furie analysis of autocorrelograms showed periodic components in spike sequences with the maxima at 3, 12, and 28 Hz. When blocking synaptic transmission with 100 mM adenosine, about 70% of the background active cells “fell silent” and the remaining 30% of neurons continued to generate action potentials. The latter seem to be actual spontaneously active neurons, i.e. they were capable of autonomous spike generation. We failed to find any correlation between the type of neuronal firing and the ability of neurons to be spontaneously active. The selective blockade of inhibitory synapses with 100 mM picrotoxine did not practically change the character of background firing activity though the

responses to stimulation became epileptic. An important conclusion to emerge from this study is that the background firing activity in cortical slices can include the actual spontaneous discharges related to intrinsic cell properties as well as those concerned with synaptic actions. Furthermore, a small number of spontaneously active neurons seem to he able to synaptically activate twice the number of cells. The inhibitory intemeurons did not significantly influence the propagation of excitation with the absence of stimulation.

In brain structures such as hippocampus, thalamus, cerebellum etc. maintained in vitro, the background activity was observed both in extracellular and intracellular recordings.2~4*2’,24~36~45*~,57*62,64 Excitatory postsynaptic potentials (EPSPs), inhibitory postsynaptic potentials (IPSPs) and action potentials (APs) occurring without any external stimulation were also found in neocortical slices.6,‘9,M,32*35~4’*56 At the same time many investigators did not reveal spontaneous spike activity in the cortical slices. This difference is probably due to peculiarities in slice preparations, variations in medium composition and probably to the changes occurring in tissue in vitro. Background activity in a slice might be based on spontaneous release of transmitter from presynaptic terminals or on the ability of some cells to endogenously generate spreading APs. One may assume that some neurons having no spontaneous activity of their own can be synaptically excited by other cells. In the latter case the activity will be the secondary one, i.e. of exogenous nature. These two groups of cells can be distinguished by synaptic blockade. Besides, the presence of inhibitory elements influencing the character of firing of the excited neurons should be taken into account. The contriACF, autocorrelation function; ACG, autocorrelogram; AP, action potential; BFA, background tiring activity; EPSP, excitatory postsynaptic potential; IPSP, inhibitory postsynaptic potential; IID, interspike interval distribution.

Abbreviations:

bution of the inhibiting neurons to the background activity can be estimated using the blockade of inhibitory synapses. In the present study we deal with temporal structure of spike trains of background active neurons in cortical slices and the influence of synaptic disconnection on them. There were three main ideas which formed the basis of our study on background firing in cortical slices: (i) obtaining statistical characteristics of spike trains to compare them with those in vivo; (ii) distinguishing between endogenous and exogenous activity by blocking all synaptic transmission with adenosine; and (iii) observing the effects of blocking inhibitory interactions through picrotoxin. EXPERIMENTAL PROCEDURES

Slice preparation

Experiments were performed on in vitro guinea-pig neocortical slices. Animals were decapitated, the cranium removed and brain dissected out and placed in cold, oxygenated artificial cerebrospinal fluid. Slices were prepared by frontal sectioning of parietal cortex block using a Teyler tissue chopper to give a thickness of 40&500 pm. Then five to 10 slices were placed at the submerged net in the superfusion chamber and sunerfused with Krebs-Ringer bicarbonate standard mediumconsisting of (in mM) 137.5 13.5 NaCl. 5 KCl. 2 C&l,. 2 MaCl,. 1.25 KHPO.. NaHCO, and 10 glucose&(pH 7.3I7.4; oxygenated wi& 95% Or, 5% CO,) and preincubated for 2 h at room temperature at a flow rate of about 3ml/min. Then the medium was gradually warmed to 37°C and after l-2 h the firing events could be observed. 915

916

i =o, 1.2 >

>n - 11, fl is the average value of ACO. % II the number of spikes in a train. and R(i) is the value 01’AC’], within an interval from AI .i to At (i + l).‘“.‘” Thcx. t!w ACF of each neuron described the neuronal firing activir) and was dependent of the number of spikes. The power spectrum calculated for each ACF made it possible to rcvcal the predominant frequencies of the regular component 111 the spike train (in the band from 0.5 to 35 Ff7).

a

RESULTS

e

300msm Fig. 1. Background patterns in neocortical slices. (a) Regular single spikes with constant interspike intervals, (b) irregular single spikes, (c) single-burst (mixed) activity, (d) bursts, (e. f) groups of spikes.

To begin with, it should be noted that neurons with background activity were very seldom found throughout the experiment. Sometimes, during a period of 3-4 h we were unable to detect the cell which could discharge for more than l-2min. At the same time some “silent” neurons responding to

a

b

Recording

Extracellular discharges were recorded with a tungsten microelectrode or a glass micropipette filled with 1 M NaCl (2-5 MD). Unit recording was performed by a highimpedance amplifier capable of at least a lOO-fold amplification and having low noise level. When firing frequency became stable (1-5 min after detection), the spike train was stored on an FM tape for further analysis. In standard medium, the duration of recordings was as long as 5 or 10 min. Sometimes, to block both excitatory and inhibitory synaptic transmission, the slices were exposed to a medium containing 100pM adenosine.” In other cases, to block only inhibitory GABAergic synapses, we added lOO/rM picrotoxine to the bathing medium.” The recordings were made before superfusion of these substances (5-lOmin), during superfusion (5-10min) and after their elimination (5-10 min). The quality of synaptic depression was tested with several monopolar stimuli (0.2 ms, 0.1 Hz). A tungsten electrode (60 kR) placed at the border of white matter was used to evoke cortical neurons orthodromically. When the discharge latency was less than l.Oms, the response was considered to be antidromic but not orthodromic. Processing neuronal activity To process spike trains, the action potentials were passed through a window discriminator unit that generated a constant-size pulse for each AP over a certain height. Statistical analysis of a spike train consisted of computation of interspike interval distribution (IID), current discharge frequency, autocorrelogram (ACG) results, and its power spectrum. The plot of firing rate shows a neuronal activity during monitoring. The ACG (or expectation density function) represents the probability for appearance of subsequent spikes after each previous one in the spike train i.e. its positive devi(within the ACG branch length), U)A3,46 ation from the average level indicates the high probability and the negative one implies the low probability for the occurrence of discharges in a definite period after spike. In the present work the ACGs were computed with At = 10 ms bin width, the ACG branch included n = 100 points, i.e. its length was n ‘At = 1000 ms. For the comparison and subsequent averaging, each ACG was normalized. The normalized autocorrelation function (ACF) was calculated using the formula: H(i) - R R(i) = N+H(O)-8’

where H(i) is the quantity with duration ranging

of interspike from AI .i

intervals of all orders to At. (i + 1) Fere,

4l-_A100

ms

200

ms

Fig. 2. Statistical characteristics of spike trains. Regularly spiking neurons had symmetrical IIDs (la) and pronounced damping oscillations in ACFs (lb). Irregular activity was characterized by the monomodal asymmetric IID (2a) and postdischarge pause in ACF (2b). Bursts and groups of mixed activity result in the initial peak in IID (3a) and in ACF (3b), though on other sites their plots do not differ from those with irregular activity. When only bursts are present in a spike train, IID and ACF look like a single peak in one or several first bins (4a, b). Groups of spikes which sometimes occur with pronounced regularity form the broader peaks in the IID. where the first mode corresponds to interspike intervals inside a group and the second to intergroup intervals (5a). the periodicity of groups is distinctly visible on the ACF (5b). In IID (a) the abscissa represents the duration of interspike intervals with 5 ms-bin width and the ordinate shows the number of intervals in the corresponding bin. In the ACF (b) the abscissa represents the time in ms (bin width is IO ma), and the ordinate shows ACF values in relative units (onset of the scale corresponds to an average level of ACF).

Firing in cortical slices

n=120

a

lt+---12

28

12

28

n=75 b ,3

/_

C

k

3

n=45

3

d ~

28

e ~ 1000

3

12

28

Hz

ms

Fig. 3. Averaged autocorrelograms (on the left) and their power spectra (on the right). The total ACF, (a) averaged over all neurons consists of two intermediate autocorrelograms ACF, and mz. In constructing ACF, (b) we snmmarized the ACF of the irregular-spiking, bursting, grouping neurons and units with mixed activity. The m, (c) was plotted by summarizing ACFs only for regularly spiking neurons. The mz, in turn, was calculated by snmming up two averaged autocorrelograms: KCF, (d) for “high-frequency” (27-28 spikes/s) and md (e) for “low frequency” (l&14 spikes/s) regularly spiking neurons. Power spectra were calculated by using Furie analysis of the corresponding intermediate average mi (be) and the total ACF, (a). The length of ACF branch was loo0 ms and the frequency in the power spectrum ranged from 0.5 to 35 Hz. The ordinates represent the relative units.

the approach of the microelectrode or to electrical stimulation of white matter were revealed in each track. Background firing activity (BFA) was mainly observed in cortical layer V and typically not earlier than after 34 h of incubation, though sometimes the BFA was found 2 h after slice preparation. In wellsurviving slices the BFA could be observed up to l&12 h after the beginning of perfusion. Statistical characteristics. In our recorded. 800-1200 Vth layer 2000 pm). according discharges discharges

experiments BFA from 128 neurons were All neurons were located at a distance of pm from the pia, i.e. they belonged to the (at the total cortical thickness of about Five kinds of BFAs can he distinguished to the pattern type: (i) regular single (55 neurons, 48%); (ii) irregular single (19, 15%); (iii) bursts (9, 7%); (iv) groups

917

(9, 7%); and (v) mixed activity representing single irregular spikes alternating with bursts or groups (36, 28%; Fig. 1). The definite shape of ACG and IID corresponded to each kind of firing activity (Fig. 2). Regular BFA was a train of single spikes with approximately equal interspike intervals. These intervals were different in different neurons. The mode (MO) and the mean (M) of IID for a regular neuron were almost the same. Their values were located within the same region: 55 < MO < 165 ms (the average value MO = 113.5 ms) and 55 < M < 165 ms (the average value m = 113.2 ms). Standard deviation (a) ranged from 4 to 46ms (5 = 12.7 ms). IIDs were symmetrical or close to symmetrical, therewith 30% of them exhibited no difference from the normal ones (P > 0.95). ACFs showed damping oscillations with the period of 55-165 ms; moreover, in 38 of 45 neurons (87%) the period was 9&165ms (Fig. 2). When bursting or grouping both IID and ACF had the initial peak. In IIDs it characterized the intraburst (4.4 + 0.9 ms) or intragroup intervals (3&50 ms; Fig. 2a,, a.,) and in ACFs it reflected the duration of bursts (24 spikes) or groups (2-5 spikes; Fig. 2b,, b4). The irregular single spikes were present in mixed activity in the addition to bursts or groups. Therefore IIDs were bimodal (Fig. 2a,). The first short peak showed the existence of bursts (or groups) and the second, the broader one, pointed to the presence of a regular component in the sequence of single spikes or bursts. The mode of the second rise of IID fell at 190-380 ms (MO = 239 ms), the mean value was 216570 ms (M = 316 ms) and standard deviation was 64-400 ms (5 = 134 ms). As a rule, between these two peaks there was a transient depression ranging from 100 to 200 ms. Sometimes a similar pause in ACF was followed by a rise of about 50ms corresponding to the short-term increase of discharge probability. When a neuron generated only single irregular APs, the IID and ACF shapes were almost the same as those for mixed activity but had no peak in the first bins (Fig. 2,). This type of activity usually remained unchanged for 10 and more minutes when the slice was not exposed to any influences. Moreover, in some cases in the process of continuous recording the regular spikes became the irregular and the irregular activity transformed into the mixed one. Frequency contents The shape of separate ACFs and their spectral characteristics varied. Therefore, in order to reveal their general typical features, ACGs were averaged. This averaging was performed in parts: first, for the neurons with grouping, bursting and mixed activities (Fig. 3b) and then for the neurons with regular firing (Fig. 3~). As regular spike trains could be high-frequency (~20 spikes/s) or low-frequency (10-14 spikes/s), the averaging of ACFs for both of them was also carried out separately (Fig. 3d, e). The first intermediate averaged correlogram (ACF,) showed an initial peak of 20 ms duration with a subsequent

S. V. KARNUP

918

a

1000

i’“”II

J

ms

,

0

1000

100

ms

J

20

101 1

30

adenosine

I

40

50

60

min

I

Fig. 4. The addition of 100 PM adenosine to the bathing solution completely suppressed regular-spiking neuronal activity. Washout of adenosine by standard solution restored firing and its characteristics almost to the initial value. On the left is the autocorrelogram (a) and its power spectrum (b) before adenosine application. The ACF and its power spectrum after washout is shown on the right. The lowest curve (c) represents changes in firing rate (bin width is 5 s). In the ACF the abscissa represents time, ms, bin width is 10 ms, and the ordinate represents relative units. On the power spectrum plot, the abscissa represents frequency, Hz, and the ordinate shows relative units.

40-ms depression to the bursting

(Fig. 3b). This peak corresponded discharges and the depression rep-

resented the low probability for the occurrence of any subsequent discharge. In the ACF, plotted for regular activity, the post-discharge depression indicated the 8&90-ms interspike intervals and corresponded to the frequency of about 12Hz (Fig. 3~). The m, constructed for the seven “high-frequency” regularly spiking neurons (Fig. 3d) showed that their interspike intervals were significantly shorter (_ 30 ms) and as a result the sharp peak at a frequency of 28 Hz was seen in the power spectrum. In ACF, of the 38 “low-frequency” regular-spiking neurons (Fig. 3e), the damping periodic oscillations were not seen, despite the fact that in each separate ACF there were five to 10 clearly marked oscillations. This was probably due to the differences in spike generation frequencies in various neurons. Therefore only the first two waves were left in the ACF, and its power spectrum showed a high peak near 12Hz. When comparing all intermediate averaged mi (Fig. 3b-e) with the final m,, computed for all neurons (Fig. 3a), it can be concluded that regularly spiking and bursting cells contributed mainly to the m,, shape. Synaptic

inactivation

For identification of endogenously and exogenously active neurons, the blocking of chemical synapses was performed by adding 100 p M adenosine

into the bath medium. The effect of synaptic dissociation on BFA was studied in spike trains recorded from 42 neurons. Adenosine application always caused a reversible disappearance of orthodromic reactions induced by electrical stimulation of surrounding tissue (Fig. 7a). At the same time antidromic responses were observed. Synaptic dissociation caused fast suppression ol BFA in 12 out of 18 regular-spiking neurons. After washing out adenosine with normal medium, the BFA character was completely restored in eight out of 12 cells falling silent (Fig. 4). When adenosine was washed out, the shape of ACF changed as compared with the initial one in the four remaining neurons. One cell began to discharge with single irregular spikes, in another bursting began and in the remaining two the firing rate significantly decreased, though discharges became regular again. It can be assumed that these 12 regular neurons were not endogenously active and discharged when affected by synaptic excitation. Another six regular neurons (33%) continued to discharge with adenosine, though their firing rates decreased. In three of them the regularity of APs was preserved (initial frequencies were 11, 25 and 35 spikes/s) and in others discharges became irregular (initial frequencies l&l7 spikes/s). After washout of adenosine the structure of firing was compleiely restored in all six cells. Addition of adenosine into the medium inhibited APs in 13 out of 19 neurons with initial irregular or

919

Firing in cortical slices

a

I_ _&a-L i 00

100

ms

100

1

bh

\k?+fyJ@~, \

C 20

i

\ ’I

0

10

20

1

I

30

I

I

50

/’ min

adenosine Fig. 5. A neuron with mixed activity (irregular single spikes and bursts) under perfusion of adenosine solution slightly decreased its discharge frequency, however the ACF and its spectrum did not change. After washout of adenosine, short-term activity exaltation was observed and then the firing rate returned to the initial level. Symbols are the same as in Fig. 4. activated ones without endogenous activity. The remaining six cells continued to discharge in adenosine solution but their frequency decreased to some extent. In one of the cells discharges became

mixed activity. The washout of adenosine restored the background activity structure in 11 cells and increased bursting in two of them. Thus, the neurons of this subgroup should be referred to synaptically

-----_

------4

E-

n

20, AL

4

I

--_-_

4000

‘S

1000

ms

p-_ Fig. 6. The addition of 100 PM picrotoxine to the medium (solid line) induced no changes either in tiring frequency (on the left) or in the shape of ACF (on the right). The addition of 100 FM adenosine (dotted line) slightly decreased the frequency of discharges. The numbers (1,2,3,4) in the plot of firing rate represent four consecutive sections of spike train for which the corresponding ACFs were computed, the abscissa represents the time with 5-s bin width, the ordinate shows the number of spikes in a bin. On the ACF abscissa represents time, ms, and the ordinate represents relative units.

920

S. V. KAKNW and intragroup intervals remained unchanged Ptcrtr toxine application never resulted in a convers-on or usual background activity to epileptic pittcrnx though the stimulus-evoked reactions were Iran+ formed into epileptiform ones (Fig. 7b). DlSCCSSlON

300

ms

Fig. 7. V-layer neuronal responses to electrical stimulation of white matter. (a) Adenosine reversibly suppressed orthodromic reactions in two different neurons. On the left, before; in the middle, during; and on the right, after perfusion with 100 PM adenosine. (b) Picrotoxine caused an epileptic discharge train which was most pronounced in response to the first stimulus in the series (0.1 pulses per second) and damped with repeated stimulation. Upper trace, a control reaction before drug application; middle and lower traces, responses during 100pM picrotoxine perfusion to the first and to the seventh stimuli. regular, in two of the cells the bursting increased and in the last three cells the character of BFA did not change. When adenosine was washed out, the ACF in four out of six neurons was the same as the initial one (Fig. 5) and for two others the ACF changed as compared with the initial one: in one cell the bursting increased and in the other, on the contrary, the single spiking increased. So, the conclusion should be made that neurons from the last subgroup (six out of 19, i.e. 3 1%) were endogenously active, however, synaptic excitation from other neurons could contribute to the activity of these units. Finally, all five neurons with initial bursting or grouping fell silent with adenosine and completely restored their firing patterns after its elimination, i.e. their activity might be of exogenous origin. It should be concluded that 12 out of 42 (29%) adenosine-tested neurons were endogenously active, i.e. capable of autonomous generation of action potentials and the remaining 30 (71%) neurons could discharge only in the presence of exciting synaptic inputs and could therefore be considered exogenously active. Inhibitory synapses blocking. In order to estimate the contribution of inhibitory units to formation of background spike train structure, we added 1OOpM of picrotoxine to the bath medium effectively blocking GABAergic synapses. In eight out of 12 neurons studied, neither firing rate nor firing structure were significantly changed (Fig. 6). In three cells the firing frequency decreased and only one cell with periodical groups showed a two-fold increase of firing rate; in the latter case periodicity of grouping significantly weakened, but group duration

In vitro intracellular investigations showed that silent neurons could generate APs when depolarizing current was passed through the microelectrode.‘h.2’ Recently, Connors and his colleagues’.“~“~4” have classified cortical neurons by their single action potential features and firing patterns during injection of intracellular current pulses. They divided the neurons observed into three physiologically distinct classes. Regularly spiking cells occurred in all layers and had a firing rate of about lo-30 single spikesis. Intrinsically bursting cells occurred only in layers IV and V and generated bursts of >3 spikes; some 01 these produced repetitive bursts. Fast-spiking cells had brief spikes. fired at high frequencies ( > 300 spikes/s) and presumably were inhibitory interneurons. Regularly spiking cells of this classification seem to be the same as our first type neurons with regular single spike trains, whereas intrinsicallybursting cells may form other kinds of the spike trains. Extracellular microelectrodes appeared to he unable to register fast-spiking cells in our experiments. Other investigators have found that depolarization of 57% of the cells resulted in tonic firing while the remaining cells (43%) had a phasic component in their responses.4’ In some neurons, further membrane depolarization resulted in the substitution of regular single spikes for bursts, but in others it resulted in the substitution of bursts for tonic sequence of regular spikes. Moreover. some cases were noted in which a cell discharged tonically at the resting potential while depolarizing or hyperpolarizing shifts induced the appearance of phase components (bursts) in its firing. Obviously. the neurons with the resting potential increasing to a critical level became spontaneously active. The oscillations nl membrane potential in stellate cells of entorhinal cortex layer II were found to be generated by a Na ’ conductance.5 Such subthreshold Na ‘-dependent rhythmic membrane oscillations as well as Ca’ ’ dependent K + conductances or voltage-dependent Ca’+ conductances’R,3R.6” indicate that rhythmicity in heterogeneous neuronal networks may be supported by different sets of intrinsic ionic mechanisms in each of the neuronal elements involved. The reasons for neuronal self-depolarization with maintenance in vitro were not specially studied. However. cell swelling is known to occur with high ~0, and low pH in the superfusion medium.” As has been mentioned, the recording of BFA usually started after 34 h of incubation. Ultramicroscopy of slices exhibits well-preserved cytoplasmic structures during 46 h and then their degradation begins;“~“““” in

Firing in cortical slices

921

inhibitory action on the transmitter release. There particular, vacuolization of a cell and its swelling are some data available on the direct postsynaptic occurs.‘**55Despite this, the electrogenous properties of neuronal membrane and conductive fibers as well influence of adenosine, where besides lowering of firing rate or blocking of the activity, the addition of as synaptic contacts are well preserved.7~‘2*‘6’8*~ Focal 100 PM adenosine led to a small hyperpolarization responses to electrical stimulation, evoked spike in two-thirds of locus coeruleus background active reactions, EPSPs and IPSPs are well reproduced neurons.” In hippocampal slices adenosine appliup to 10-12 h from the beginning of incubation.‘6*” cations resulted in marked hyperpolarization of It is possible that at prolonged survival the changed CA1 pyramidal neurons. 23s2 The authors showed neurons are spontaneously active. These neurons, that this effect was due to an increase in two potassin turn, can produce synaptic effects on the other cells and provoke the occurrence of EPSP and ium conductances (Ca2+-dependent and Ca2+sometimes action potentials in them. On the other and voltage-independent) and was mediated through A, receptors. In our experiments a great number hand, we occasionally detected background spike trains as early as 2 h from the beginning of the of neurons stopped discharging with adenosine experiment. application, and the cells which continued to No matter what caused spontaneous discharges, generate APs often decreased their firing rate. the excitation leads to release of transmitter into This fact might be accounted for in two ways. On the synaptic gap and induces the occurrence of the one hand, it could be due to depression of the ability to spike generation in some neurons with EPSP or IPSP. For instance, intracellular recordings from layer II entorhinal cortex slices from postnatal the decrease of membrane potential. In this case, all the neurons observed are considered to be sponrats showed a pronounced spontaneous synaptic taneously active and independent of synaptic inputs. activity which could create large depolarizing events and give rise to bursting.” Perfusion with On the other hand, if the main cause of discharge the N-methyl-D-aspartate receptor antagonist 2- disappearance was the synaptic inactivation, those amino-S-phosphonovaleriate abolished the spon-70% of neurons which discharged only in the taneous activity. This fact implies that the activity presence of synaptically exciting input did not exhibit their own spontaneous activity. On the contrary, the observed may result from spontaneously released transmitter interacting with a large population of remaining -30% of the cells with background postsynaptic N-methyl-D-aspartate receptors. activity were capable of autonomous (spontaneous) The decrease of Ca2+ concentration to 0.245 mM pulsation and could create some level of synaptic and the increase of Mg2+ (to 8 mM), or the addition excitation. of Co2+ or Mn2+ into the bath medium is the convenIn BFA of neurons in vim the postdischarge pause tional way to block synaptic transmission.2s~29@‘@’ of 10-500 ms duration is usually interpreted as an However, firstly such a change in ion composition of inhibiting pause resulting from recurrent inhibition.34*43,58@,69 Since similar pauses are revealed the medium affects not only presynaptic membranes, making transmitter release difficult, but it also under autocorrelation analysis of BFA in vitro, it influences the somatic membrane and modulates might be suggested that the same recurrent inhibition its excitation.3 Secondly, it becomes impossible to mechanism responsible for the periodicity of study effects of Ca2+ concentration changes using cell discharges takes place here. However, in such techniques of synaptic blockade. Therefore, we this case inactivation of inhibitory synapses have used micromolar concentrations of adenosine should lead to the disappearance of long postdissolved in standard medium for blocking synaptic discharge pauses or at least to their decrease. The transmission. Low doses of purine nucleotides, use of picrotoxine for this purpose induced adenosine in particular, are known to be capable of no changes in spike trains to say nothing of “disinactivating the chemical synapses;‘5”7*44~48~53 in this inhibition”. Therefore, low probability of spike ocway the adenosine-induced reduction of transmitter currence during 50-200ms after the previous release is observed. ‘3~‘4Adenosine application lowers single spike or burst was not due to the activity of the amplitude of field potentials at concentrations inhibitory interneurons. The lack of spontaneous of O.l-S.OpM, and at 100pM it almost completely epileptic discharges also points to the fact that, under suppresses them.‘5~39~5’These concentrations of normal conditions, the interneurons do not produce adenosine with orthodromic stimulation prevent any inhibitory influence on the surrounding cells. the occurrence of both EPSPs and IPSPS.~~,~~ In the studies of epileptogenesis on neocortical Adenosine and its derivatives are considered to slices, the application of GABA antagonists penicillin affect the A, receptors4’s6’ which are involved and bicuculline induced neither spontaneous memin Ca2+ re-uptake. 67 On the other hand, there brane potential depolarizing shifts nor convulsive is evidence that inhibition of transmission in the discharges, though in response to stimulation EPSPs lateral olfactory tract-pyramidal cell synapses by of short latency and the depolarizing shifts with adenosine is mediated by receptors of the A, catlonger latencies were observed on the background egory. ‘o*39 Thus, the inhibitory effects of purines on of which high-frequency paroxysmal discharges appear.22s49,66 neuronal activity are attributed to a presynaptic

s. v.

922

There is reason to believe that the formation of the background spike train in neurons may reflect either the peculiarities of its interactions with other excitatory units in the slice or its

KARNUP

intrinsic functional state. In some neurons mrfumai synaptic excitatory input from a small number 4 spontaneous neurons may be sufficient for spike initiation.

REFERENCES

1. Agmon A. and Connors B. W. (1989) Repetitive burst-firing neurons in the deep layers of mouse somatosensory L,crrtex. Neurosci. Left. 99, 137-141.

2. Andersen P., Bland B., Skrede K., Sveen 0. and Westgaard R. (1972) Single discharges in brain slice maintained ;n vitro. Acta physiol. stand. 84, IA-2A. 3. Alger B. E., Dhanjal S. S., Dingledine R., Garthwaite J., Henderson G., King G. L., Lipton P., North A., Schwartzkroin P. A., Sears T. A., Segal M., Wittingham T. S. and Williams J. (1984) Brain Slice Methods. In Brain Slicr.r (ed. Dingledine R.), pp. 381437. Plenum Press, New York. 4. Alger B. E. and Nicoll R. A. (1980) Spontaneous inhibitory postsynaptic potentials in hippocampus: mechanisms for tonic inhibition. Brain Res. 200, 195-200. 5. Alonso A. and Llinas R. (1989) Subthreshold Na-dependent theta-like rhythmicity in stellate cells of entorhinal cortex layer II. Nature 342, 1755177. 6. Angel A. and Pollard C. E. (1985) A comparison of the effect of uretane on spontaneously active cells in the rat somatosensory cortex in vivo and in vitro. J. Physiol. 364, 33P. 7. Bak I. J., Misgeld V., Weiler M. and Morgan E. (1980) The preservation of nerve cells in rat neostriatal slices maintained in vitro. A morphological study. Brain Res. 197, 341-353: 8. Bingmann D., Kold G. and Speckmann E.-I. (1982) Effects of elevated ~0, values in the superfusate on the neuronal activity in hippocampal slices. In Physiology and Pharmacology of Epileptogenic Phenomena (ed. Klee M. R. PI al.). pp. 97-104. Raven Press, New York. 9. Chagnac-Amitai Y. and Connors B. W. (1989) Synchronized excitation and inhibition driven by intrinsically bursting neurons in neocortex. J. Neurophysiol. 62, 1149-l 161. 10. Collins G. G. S. and Anson J. (1985) Adenosine A, receptors mediate the inhibitory effects of exogenous adenosine in the rat olfactory cortex slice. Neuropharmacology 24, 1077-1084. 11. Connors B. W. and Gutnic M. J. (1990) Intrinsic firing patterns of diverse neocortical neurons. Trends Neurosct. 13, 99-104.

12. Crepe1 F., Dhanjal S. S. and Garthwaite J. (1981) Morphological and electrophysiological characteristics of rat cerebellar slices maintained in vitro. J. Physiol. 316, 127-138. 13. Dolphin A. C. and Prestwich S. A. (1985) Pertussis toxin reverses adenosine inhibition of neuronal glutamate release. Nature 316, 1488150.

14. Dunwiddie T. V. and Haas H. L. (1985) Adenosine increases synaptic facilitation in the in vitro rat hippocampus: evidence for a presynaptic site of action. J. Physiol. 369, 3655377. 15. Dunwiddie T. V. and Hoffer B. J. (1980) Adenine nucleotides and synaptic transmission in the in vitro rat hippocampus. Br. J. Pharmac. 69, 59-68.

16. Franc G. (1972) Brain Slices. In The Structure and Function of Nervous Tissue (ed. Bourne G. H.), Vol. 6, pp. 417-465. Academic Press, New York. 17. Frotscher M. and Misgeld V. (1989) Characterization of input synapses on intracellularly stained neurons in hippocampal slices: an HRP/EM study. Expl Brain Res. 75, 327-334. 18. Frotscher M., Misgeld V. and Mitch C. (1981) Ultrastructure of mossy fiber endings in vitro hippocampal slices Expl Brain Res. 41, 247-255.

19. Galvan M., Franz P. and Constanti A. (1985) Spontaneous inhibitory postsynaptic potentials in guinea-pig neocortex and olfactory cortex neurons. Neurosci. Lett. 57, 131-135. 20. Gilette M. U. (1985) Preparation of brain slices from the suprachiasmatic nuclei of rat can reset the circadian clock. J. Physiol. 369, 55P.

21. Gerstein G. L. and Kiang Y. S. (1960) An approach to the quantitative analysis of electrophysiological data from single neurons. Biophys. J. 1, 15-38. 22. Gutnic M. J., Connors B. W. and Prince D. A. (1982) Mechanisms of neocortical epileptogenesis in vitro. J. Neurophysiol. 48, 1321-1335.

23. Haas H. L. and Greene R. W. (1988) Electrophysiological analysis of effects of exogenous and endogenous adenosine in hippocampal slices. In Neurotransmitters and Cortical Function. From Molecules to Mind (eds Avoli M., Reader T. A.. Dykes R. W. and Gloor P.), pp. 483494. Plenum Press, New York. 24. Habilitz J. J. and Johnston D. (1981) Endogenous nature of spontaneous bursting in hippocampal pyramidal neurons. Cell. molec. Neurobiol. 1, 3255334. 25. Hotson J. R. and Prince D. A. (1980) A calcium-activated hyperpolarization follows repetitive firing in hippocampal neurons. J. Neurophysiol. 43, 409419. 26. Jahnsen H. (1986) Electrophysiological characteristics of neurons in the guinea-pig deep cerebellar nuclei in vitro. J. Physiol. 372, 129-147.

21. Jahnsen H. and Llinas R. (1984) Electrophysiological

properties of guinea-pig thalamic neurons: an in vitro study J.

Physiol. 349, 205-226.

28 Jahnsen H. and Llinas R. (1984) Ionic basic for the electro-responsiveness and oscillatory properties of guinea-pig thalamic neurones in vitro. J. Physiol. 349, 221-247. 29 Jefferys J. G. R. and Haas H. L. (1982) Synchronized bursting of CA1 hippocampal pyramidal cells in the absence of synaptic transmission. Nature 300, 4488450. 30 Jones R. S. G. and Heinemann U. (1989) Spontaneous activity mediated by NMDA receptors in immature rat entorhinal cortex in vitro. Neurosci. Lett. 104, 93-98. 31. Karnup S. V. (1989) Background neuronal activity with blockade of synaptic inhibition in guinea-pig neocortical slices. Neurophysiologv 21, 13- 19.

Firing in cortical slices

923

32. Kamup S. V., Bortnic A. T. and Zhadin M. N. (1985) Structure of background neuronal activity in thin slices of guinea-pig neocortex. Neurophysiology 17, 441449. 33. Kamup S. V. and Kolomoets A. V. (1987) Neuronal firing in guinea-pig neocortical slices during synaptic blockade of transmission by adenosine. Neurophysiology 19, W-606. 34. Kamup S. V. and Zhadin M. N. (1980) Autocormlation analysis of background activity of cortical neurones in learning. Zh. vyssh. nerv. Deyat. J. PT P&n-3@ 738-746 (in Russian). 35. Kato H., Ito Z., Matsuoka S. and Sacurai Y. (1973) Electrical activities of neurones in the sliced human cortex in v&o. Electroenceph. din. Neurophysiol. 35, 457-462. 36. Kobayashi S. (1986) Warm- and cold-sensitive neurons inactive at normal core temperature in rat hypothalamic slices. Brain Res. 362, 132-139. 37. Lee K. and Schubert P. (1982) Modulation of an inhibitory circuit by adenosine and AMP in the hippocampus. Brain Res. 246, 311-314. 38. Llinas R. R. (1988) The intrinsic electrophysiological properties of mammalian neurons: insights into central nervous

system function. Science. 242, 1654-1664. 39. McCabe J. and Scholfield C. N. (1985) Adenosine-induced depression of synaptic transmission in the isolated olfactory cortex: receptor identification. PJugers Arch. ges. Physiol. 403, 141-145. 40. McCormik D. A., Connors B. W., Lighthall J. W. and Prince D. A. (1985) Comparative eiectrophysiology of pyramidal and sparsely spiny stellate neurons of the neocortex. J. Neurophysiol. 54, 782-805. 41. McIntyre D. C. and Wong R. K. S. (1986) Cellular and synaptic properties of amygdala-kindled pyrifotm cortex in vitro. J. Neurophysiol. 55, 1295-1307. 42. Montro R. J., Lopes-Bameo J. and Jassik-Gerschenfield D. (1988) Differential burst firing modes in neurons of the mammalian visual cortex in vitro. Brain Res. 460, 168-172. 43. Moore G. P., Perkel D. H. and Segundo J. P. (1966) Statistical analysis and functional interpretation of neuronal spike data. A. Rev. Physiol. 28, 493-522. 44. Okada G. and Oxawa S. (1980) Inhibitory action of adenosine on synaptic transmission in the hippocampus of the guinea-pig in vitro Eur. J. Pharmac. 18, 483492. 45. Okamoto K., Kimura H. and Sakai Y. (1984) Miniature synaptic potentials recorded intracellularly from Purkinje cell dendrites in guinea-pig cerebellar slices. Brain Res. 311, 281-287. 36. Perkel D. H., Gerstein G. L. and Moore G. P. (1967) Neuronal spike trains and stochastic point process. I: The single spike trains. Biophys. J. 7, 391418. 47. Phillis J. W. (1982) Evidence for an A-like adenosine receptor on cerebral cortical neurones. J. Pharm. Pharmac. 34, 453-454. 48. Phillis J. W., Edstrom J. P., Kostopulos G. K. and Kirkpatrik J. R. (1979) Effects of adenosine and adenine nucleotides on synaptic transmission in cerebral cortex. Can. J. Physiol. Pharmac. 57, 128+1312. 49. Prince D. A. (1978) Neurophysiology of epilepsy. A. Rev. Neurosci. 1, 395-415. 50. Randle J. C. R., Bourque C. W. and Renaud L. P. (1986) Characterization of spontaneous and evoked inhibitory postsynaptic potentials in rat supraoptic neurosecretory neurons in vitro. J. Neurophysiol. 56, 1703-1717. 51. Sanderson G. and Scholfield C. N. (1986) Effects of adenosine uptake blockers and adenosine on evoked potentials of guinea-pig olfactory cortex. PpUgers Arch. ges. Physiol. 406, 25-30. 52. Segal M. (1982) Intracellular analysis of postsynaptic action of adenosine in the rat hippocampus. Eur. J. Pharmac. 79, 193-200. 53. Scholfield C. H. (1986) Evidence for adenosine blockade of synaptic transmission by enhancement of presynaptic K+-conductance in guinea-pig brain slices. J. Physiol. 371, 63P. 54. Schurr A., Reid K. H., Tseng M. T. and Edmonds H. L. (1984) The stability of hippocampal slice preparation: an electrophysio-logical and ultrastructural analysis. Brain Z&s. 297, 357-362. 55. Schurr A., Reid K. H., Tseng M. T., Edmonds H. L. and Rigor B. M. (1985) A dual chamber for comparative studies using the brain slice preparation. Comp. Biochem Physiol. &2A, 701-704. 56. Shaw C. and Teyler T. J. (1982) The neural circuitry of the neocortex examined in the in vitro brain slice preparation. Brain Res. 243, 35-47. 57. Shefner S. A. and Chiu T. H. (1986) Adenosine inhibits locus coeruleus neurons: an intracellular study in rat brain slice preparation. Brain. Res. 366, 364-368. 58. Shosaku A. (1986) Cross-correlation analysis of a recurrent inhibitory circuit in the rat thalamus. J. Neurophysiof. 55, 103cHO43. 59. Siggins G. R. and Schubert P. (1981) Adenosine depression of hippocampal neurones in vitro: an intracehular study of dose-dependent actions on synaptic and membrane potentials. Neurosci. ktt. 23, 55-60. 60. Snow R. W. and Dudec F. E. (1984) Synchronous epileptiform burst without chemical transmission in CAI, CA3 and dentate areas of the hippocampus. Brain Res. 298, 382-385. 61. Stone T. W. (1982) Cell membrane receptors for purines. Biosci. Rep. 2, 77-90. 62. Thomson A. M. (1984) Slow, regular discharge in suprachiasmatic neurones is calcium-dependent in slices of rat brain.

Neuroscience 13, 761-767. 63. Traub R. D., Miles R. and Wong R. K. S. (1989) Model of the origin of rhythmic population oscillations in the hippocampal slice. Science 243, 1319-1325. 64. Trulson M. E., Howell G. A., Brandstetter J. M., Frederickson M. H. and Frederickson C. J. (1982) In vitro recording of raphe unit activity: evidence for endogenous rhythms in presumed serotoninergic neurons. Life Sci. 31, 785790. 65. Tsau Y. and Chen P.-X. (1989) Normalized power spectrum density function analysis on spike trains II. Rhythms of unit activities in the somatosensory cortex of cats. Znt. J. Neurosci. 49, 123-131. 66. Voskuyl R. A. and Ter Keurs H. E. D. J. (1978) Excitatory increase of neurons in olfactory cortex slices of the guinea-pig after penicillin administration. Brain Z&s. 156, 83-96. 67. Wu P. H., Phillis J. W. and Thierry D. L. (1982) Adenosine receptor agonists inhibit potassium evoked calcium uptake by rat brain cortical synaptosomes. J. Neurochem. 39, 700-708. 68. Yamamoto C., Buk I. J. and Kurokawa M. (1970) Ultrastructural changes associated with reversible and irreversible suppression of elecrical activity in olfactory cortex slices. Expl Brain Res. 11, 36&372.

924

S. V. KARNUP

69. Yamamoto M. and Nakahama H. (1986) Time series analysis of neuronai impulse signal and its physic:iogrc.r; significance J. Physioi. Sot. Japan. 48,491-504. 70. Zhadin M. N. (1988) On normalization of the cross-correlation function of pulse activity. Studio hioph~,.tic~,. 1%. 95-103.