Developmental Brain Research 123 (2000) 151–164 www.elsevier.com / locate / bres Research report Activity-dependent development of spontaneous bioelectric activity in organotypic cultures of rat occipital cortex ´ 1 , Klaus Albus* Diego Echevarrıa ¨ , Germany Department of Neurobiology /192, Max Planck Institute for Biophysical Chemistry, P.O. Box 2841, D-37070 Gottingen Accepted 2 August 2000 Abstract The development of spontaneous bioelectric activity (SBA) in organotypic tissue cultures (OTCs) from rat occipital cortex was studied by means of extracellular recording techniques in OTCs grown normally for 6–51 days in vitro (DIV), and in OTCs in which SBA had been silenced from DIV 4 on for 2 to 3 weeks by elevating the Mg 21 levels in the growth medium. The proportions of spontaneously active neurones increased from about 25% at 6–14 DIV to more than 80% beyond the third week in vitro. Mature neurones discharged at shorter intervals and more vigorously than immature neurones; the developmental increase in firing rate was not significant, however. In OTCs 6–14 DIV the majority of spontaneously active neurones fired sluggishly in a regular manner. The remaining neurones fired action potentials in the form of discrete bursts resembling interictal activity in vivo. The proportions of these neurones increased from about 40% at 6–14 DIV to more than 80% beyond the third week in vitro. During development in vitro the mean burst duration increased from 3.5 s to about 8 s whereas the mean burst rate (between 0.7–1 bursts / min) remained constant. Activity-deprived neurons had low firing rates and fired action potentials in the form of discrete bursts with a mean burst rate of 0.4 / min. The proportions of spontaneously active neurons, the variability of neuronal firing and the viability of the explants either were not altered by the activity blockade or had recovered to control values after 5–6 days in normal growth medium. We conclude that in OTCs of rat neocortex the absence of SBA during development in vitro delays the maturation of excitatory mechanisms responsible for the developmental increase in firing intensity. The development of burst firing modes is less affected by activity blockade.  2000 Elsevier Science B.V. All rights reserved. Theme: Development and regeneration Topic: Cerebral cortex and limbic system Keywords: Neocortex; Spontaneous action potential; Extracellular recording; Explant culture; Activity blockade 1. Introduction The development of structure and function of the mammalian visual system is conditional upon the presence of spontaneous bioelectrical activity. Continuously silencing retinal activity blocks the structural development of retinogeniculate synapses [31], delays the growth of *Corresponding author. Tel.: 149-551-201-1652; fax: 149-551-2011788. E-mail address: kalbus@gwdg.de (K. Albus). 1 Present address: Department of Morphological Sciences, Faculty of Medicine, 30071 Murcia, Spain. geniculocortical axon arbors [2] and of synaptogenesis in the visual cortex [47] and interferes with the development of ocular domains in the superior colliculus [51] and the visual cortex [50]. Even before the onset of visual experience, spontaneous retinal activity seems to be necessary for the formation of highly stereotyped patterns of connections in the lateral geniculate nucleus (LGN) [40]. Impulse blockade in the LGN has been shown to retard the developmental expression of a broad range of receptors in the developing visual cortex [26]; intracranial infusion of the sodium channel blocker tetrodotoxin (TTX) impairs the initial targeting made by thalamic axons when projecting to the visual cortex [12]. The effect of impulse blockade in central regions of the visual system on the development of bioelectric activity in these structures has not been investigated so far in vivo. 0165-3806 / 00 / $ – see front matter  2000 Elsevier Science B.V. All rights reserved. PII: S0165-3806( 00 )00089-4 152 ´ , K. Albus / Developmental Brain Research 123 (2000) 151 – 164 D. Echevarrıa Blocking neural activity in the retina is not a means of exploring the activity-dependent development of bioelectric activity in the visual cortex since total visual deafferentation does not block spontaneous activity in visual cortical neurons [33]; the remaining activity seems to guarantee a normal development of visual responsiveness of cortical neurons even after longer periods of binocular impulse blockade [50]. Thus, in order to study the activitydependent development of structure and function in central regions of the nervous system, long-term in vitro systems like dissociated cell cultures or organotypic cultures (OTCs) have been employed. Neurons in explant cultures of spinal cord and neocortex generate spontaneous bioelectric activity (SBA) in the absence of stimulation [16,13]. Crain et al. [17] demonstrated that characteristic spontaneous and evoked synaptic network discharges could be recorded from embryonic mouse cerebral and spinal cord tissues explanted at presynaptic stages and maintained for many weeks in culture media containing impulse-blocking concentrations of xylocaine or high Mg 21 prior to return to control medium (for review see Ref. [20]). Others have demonstrated that the development of SBA is strongly influenced by impulse blockade starting shortly after explantation or cell seeding. The directions of the effects observed are not however, uniform and seem to depend on the type of culture and the brain region cultivated. SBA is abnormally low in OTCs of spinal cord which had been exposed to TTX [5,6]. In contrast, following the cessation of TTX-treatment a prolonged increase in electrical excitability of the neuronal networks in OTCs of cerebellum [49] and dissociated cell cultures from neocortex [14,42] occurs. The increases in mean firing rate of single neurons and in proportions of neurons with stereotyped bursting have been accounted to a decrease in GABA neurotransmission [45]. The suggestion of an increase in the ratio of excitatory to inhibitory mechanisms in activity-deprived cultures is corroborated by findings that TTX treatment of OTCs reduces GAD-mRNA expression (glutamic acid decarboxylase; the synthetic enzyme for GABA) in neocortical interneurones [46] and significantly reduces the number of somatic inhibitory synapses on Purkinje cells [49]. The increased frequency of calcium transients in activity-deprived dissociated cell cultures from neocortex also indicates hyperactivity [38]. We have investigated the development of SBA in OTCs from rat occipital cortex in normally growing OTCs and in OTCs in which SBA was continuously blocked for 2–3 weeks. We show that activity blockade significantly reduces the level of SBA in neocortical OTCs and in addition, alters the temporal organization of spike trains. Thus, OTCs from neocortex react differently to activity blockade than dissociated cell cultures from neocortex or OTCs from cerebellum. Parts of the findings have been published in abstract form [22]. 2. Materials and methods 2.1. Preparation of organotypic tissue cultures Organotypic brain cultures were prepared and maintained in vitro using the roller-tube technique introduced by Hogue [29] and later modified by Costero and Pomerat ¨ [15] employing the protocol given by Gahwiler [25]. Briefly, rats (Wistar strain) 2 or 3 days old were anaesthetised by lowering their body temperature on ice and decapitated, and one cortical hemisphere was removed. The cortex was dissected according to the atlas of Paxinos et al. [39]; visual areas in very young rats cannot be distinguished histologically. Blocks of occipital cortex were cut parasagittally into 350-mm-thick slices on a McIlwain tissue chopper. Slices were placed on a Petri dish with Gey’s balanced salt solution (GBSS) supplemented with D-glucose to a final concentration of 0.65% at 48C for a 30–60-min recovery period. Individual slices were then placed on cleaned coverslips (11324 mm), fixed by a drop of chicken plasma coagulated with thrombin (20 ml each) and placed into reagent tubes supplied with 0.75 ml semi-artificial medium (50% Eagle’s basal medium, 25% Hank’s balanced salt solution (HBSS), 25% of inactivated horse serum, 1 mM final concentration of L-glutamine, 0.65% final concentration of D-glucose and 1.26 mM final concentration Ca 21 ). All procedures were done in sterile conditions. Slices were incubated at 368C in dry air up to 12 weeks in a roller tube drum (10 rev. / h). After 2–3 days an antimitotic cocktail (5-fluoro-2-desoxyuridine, cytosine-bD-arabinofuranoside and uridine; each at 4 mM final concentration) was added to the nutrient medium for 24 h. Thereafter, medium was changed every 3–4 days. Gross histological structure of the slices was revealed by Nissl staining. 2.2. Data acquisition and evaluation Cultures from 6 to 51 DIV were placed in a recording chamber and continuously superfused with HBSS at a rate of 25 ml / h (bath temperature 36618C); in all experiments, the final concentration of Ca 21 in the superfusion medium was increased to 3.1 mM. The pH in the recording chamber was adjusted to the pH in the test tube in which the explant was grown and was held constant within 60.2. In addition, the air in the recording chamber was saturated with oxygen. Extracellular recordings were performed with glass capillaries (1.5-mm inner tip diameter) filled with superfusion medium. Spike trains of spontaneously active neurones were stored on magnetic tape and analysed off-line using software commercially available (Brain-Wave) or developed by Dr. Gras (R-Time; Dept. of Zoology, Universi¨ ty Gottingen). Neuronal activity was analysed by calculating interspike interval (ISI) distributions and autocorrelog- ´ , K. Albus / Developmental Brain Research 123 (2000) 151 – 164 D. Echevarrıa rams (ACG). The following parameters were measured for each spike train: firing rate (number of spikes / s), median value and modal value of ISI distributions, the position (latency) of peaks in ISI histograms, width and latency of peaks in ACG histograms, duration and rate (number / min) of grouped spike discharges (denoted in the following as bursts; see Ref. [28]) and burst ratio. Burst duration was measured either directly from the spike time histograms (STHs; binwidth 1 s) or from ACG histograms (binwidth 50 ms). The burst ratio was determined by calculating the number of intervals shorter than the mean interval as a proportion of all interspike intervals, following the procedure proposed by Habets et al. [28] (see also Ref. [42]). In addition to burst ratio we note the burst ratio standardised with respect to firing rate. The variability of spontaneous activity over time was determined by calculating the coefficient of variation of the firing rate in a sliding window 120 s long (CV120s) which was moved in time steps of 60 s. ACGs were subjected to a x 2 -test for identity with a homogeneous distribution. ACGs with peaks significantly above the expected value for a homogeneous distribution (P value ,0.001) were differentiated from those without significant peaks. Single units were classified into two types: non-spontaneously active neurones (NonSp-neurones) and spontaneously active neurones (Sp-neurones). Sp-neurones with a burst ratio ,5 were denoted as spontaneously nonbursting neurones (SpnonB-neurones) and Sp-neurones with a burst ratio $5 as spontaneously active bursting neurones (SpBneurones). The proportion of Sp-neurones was calculated on the basis of all neurones recorded per developmental stage. Proportions of SpB-neurones per developmental stage were calculated considering only neurones which could be recorded for at least 4 min. NonSp-neurones were detected by electrically stimulating the gray or white matter at various distances from the recording site with a glass insulated tungsten electrode (stimulation frequency 0.2 Hz, pulse width 50 ms). Neurones firing spontaneously five action potentials or less within a period of 10 min (N587) were classified as rarely spontaneous active neurones and included into the group of NonSp-neurones. A total of 19 single units with modal intervals of $5000 ms (N58) and / or with a median value in the ISI distribution exceeding 20 s (N511), which were present at all developmental stages, were accepted as spontaneously active but were excluded from calculation of mean values and medians. The normal development of spontaneous bioelectric activity (SBA) in vitro was studied in 41 OTCs; 39 OTCs were explanted at PND 2, and two OTCs at PND 3. From these 41 OTCs 579 neurones were recorded. For quantitative analysis we considered only SP-neurons recorded longer than 4 min (N5260). The experiments were run longitudinally: OTCs from one hemisphere (batch) were used at different developmental stages. The number of 153 OTCs from the same batch contributing to one stage was #2 (6–7 DIV, 15–21 DIV, 22–28 DIV, 29–35 DIV) or #3 (8–14 DIV, .35 DIV). Analysis of variance (see below) of the intra-stage data demonstrated no difference between the OTCs contributing to one stage. The few exceptions found (P value ,0.05; less than 2% of all comparisons) did not contribute to inter-stage differences if present. We therefore, considered the data points at different developmental stages to be independent. In another set of experiments SBA was chronically blocked in OTCs for 14–23 days by elevating the Mg 21 levels in the growth medium; blockade started at DIV 4. The blockade is due to the lowered Ca 21 / Mg 21 ratio which blocks synaptic transmission [21,43]. High levels of Mg 21 do not necessarily block spontaneous discharges in cerebellar explant cultures [24,49,53]. We therefore, tested the effectiveness of 5 mM and 11 mM Mg 21 , respectively, in acute experiments in three explants (22–28 DIV, 65 neurones). The chronic experiments were performed in 14 OTCs which had been returned to normal medium for 0–6 days. Since the normal development study had revealed that the increase in the proportion of Sp-neurones and the improvement of firing properties were prominent at the end of the third postnatal week, the activity blockade was induced at 4 DIV and continued at least until DIV 18 and at the most until DIV 27. After that time, OTCs were returned to normal medium and explored with microelectrodes. From the 285 neurones recorded, 76 Sp-neurones were quantitatively analysed as described above. Recordings at recovery day 0 (Rec(0)) started 3 h after returning to normal Mg 21 levels. As controls (control (s.h.)) four OTCs (65 neurones; 52 quantitatively analysed) from the hemispheres from which the explants for the chronic activity blockade experiments were taken, were grown under normal conditions. The data from these controls were also considered for the analysis of normal development. Differences in mean values between developmental stages were calculated by using the Kruskal Wallis test with Dunn’s multiple comparison test as post test. Differences were accepted as statistically significant if the P value (two-tailed) was less than 0.05 (GraphPad Prism, V 2.0). 3. Results 3.1. Normal development of SBA in vitro 3.1.1. Active sites and firing variability The experimental approach consisted in searching for spontaneous and / or electrically evoked activity of single units by moving the electrode through the tissue in a direction nearly perpendicular to the surface. If a neurone was detected the site was denoted as an active site. In the 154 ´ , K. Albus / Developmental Brain Research 123 (2000) 151 – 164 D. Echevarrıa normal development study at least 14 sites per OTC were explored (mean6S.D.518.862.7); an OTC was considered for evaluation only, if at least nine neurones could be recorded. The mean number of neurones recorded / OTC was 14.264.5; the proportions of active sites / sites explored per OTC range between 20% and 100% and are on average 75.6616.7% (mean6S.D.; Table 1). A statistical analysis demonstrates that OTCs recorded from shortly after explantation have about as many active sites as OTCs at later developmental stages. The variability of SBA over time was determined for each Sp-neurone by calculating the coefficient of variation of the firing rate (CV120s; Table 1). The CV120s ranges from 11.68 to 204.3; the range of single values and the median of the distributions do not change during development in vitro. Thus, the neural populations recorded at different developmental stages are homogeneous with respect to the parameter firing variability over time. 3.1.2. Developmental changes in firing strength and temporal structure of spike trains The main developmental events during time in vitro are increases in the proportions of Sp-neurones and SpBneurones and increases in the firing intensity of Sp-neurones. During the first and second week in vitro only about 25% of the neurones are spontaneously active (Table 1). Spike trains characteristic of these early developmental stages are reproduced in Fig. 1. About 60% of the Spneurones fire sluggishly at a rate of usually less than 0.2 Hz (Fig. 1A–C) without any clear indication of discrete bursts (Fig. 1C9). The remaining neurones fire action potentials in the form of short discrete bursts, at rather long intervals during which either irregular spontaneous activity (Fig. 1D) or no spontaneous activity (Fig. 1E) occurs. Thus, 37.5% of these neurones were classified as SpBneurones at 6–7 DIV, and 18.7% at 8–14 DIV. Bursts last from about 200 ms to s; the intraburst firing rate seldomly exceeds 20 Hz. In 80% of the neurones one type of burst is repeated at similar intervals with rather constant duration and intraburst firing rate (Fig. 1E). In the remaining 20% the repetition rate is irregular or two different types of bursts are produced (see below, Fig. 2). The proportions of Sp-neurones increase to about 60% in the third week and to more than 80% at later stages of development in vitro. At later developmental stages most of the Sp-neurones are classified as SpB-neurones (Table 1); accordingly, the proportions of neurones displaying regular nonburst activities drop below 20%. Beyond the third week in vitro the majority of neurones now fire in the form of discrete bursts at rather long intervals. Bursts last on average longer (7–8 s) than during the first 2 weeks in vitro (3–4 s; see Table 5) whereas the burst rate has not changed (Table 2). Typical spike trains recorded in OTCs 3 weeks and longer in vitro are presented in Fig. 2. Often a burst continues over 10 s; in a few cases (Fig. 2C and E) bursts last up to 30 s. The intraburst firing rate often exceeds 20 Hz. Bursts have either a predominantly tonic (Fig. 2B, B9; D, D9; E, E9) or phasic-tonic (Fig. 2A, A9; C, C9) appearance; tonic discharges may be interrupted by short silent intervals (Fig. 2B, B9; C, C9; E, E9). The neurone in Fig. 2E9 in fact, fires repetitive short bursts at a rate of |3 Hz (see also Figs. 5 and 6). In more than 80% of these neurones stereotyped burst patterns consisting of one, or more than one burst type are seen. For example, three of the four discharge sequences of the neurone shown in Fig. 2C consist of an initial burst which after a short silent period is followed by a long tonic discharge lasting up to 30 s. The discharge sequence of another neurone (Fig. 2D) starts with an initial burst lasting about 5 s which after a short pause is followed by a 10–15 s long burst and subsequently by a number (N58–10) of shorter bursts; this pattern is repeated with minor variations. The development of SBA in individual batches of OTCs (i.e. OTCs explanted from one hemisphere) with respect to the time course of development, and to proportions of Sp-neurones (Fig. 3A–D) matches closely the mean developmental trend shown in Table 1; i.e. interbatch variability is generally low. The delayed increase in the proportion of Sp-neurones in one OTC as shown in Fig. 3F was an exception. 3.1.3. Quantitative evaluation of developmental changes in SBA SBA was quantified by the firing rate, the median and the modal value of the ISI distribution, the burst ratio, the standardised burst ratio and the burst rate (Table 2). Two different developmental trends are apparent. From the first to the second week in vitro the firing intensity decreases Table 1 Development of SBA in OTCs of rat occipital cortex. Summary of methodical and basic experimental data. Average recording time (all neurons) was 11.463.4 min (mean6S.D.) Days in vitro 6–7 8–14 15–21 22–28 29–35 .35 No. batches No. OTCs No. neurons Active sites (%; mean6S.D.) CV120s (mean6S.D.) Sp-neurons (%) SpB-neurons (%) 5 7 87 67.3624.5 49.8621.5 26.4 37.5 3 8 112 83.2613.5 58.2643.3 25 18.7 5 8 119 80.2619.5 51.2636.7 59.7 59.6 4 8 98 66.5615.1 62.4637.5 87.7 81.5 4 5 88 7767.3 67.1635.5 85.2 89.5 2 5 75 77.962.8 55.4631.4 82.7 81.8 ´ , K. Albus / Developmental Brain Research 123 (2000) 151 – 164 D. Echevarrıa Fig. 1. Spontaneous activity of single neurons in OTCs of rat occipital cortex 6–14 DIV. (A–E) Spike time histograms (binwidth 1 s, duration 10 min). Underlined segments of C, D and E are shown at an enlarged time scale in C9, D9 and E9 (binwidth 40 ms, duration 25 s). The scale at the left hand side of each histogram indicates the number of action potentials / bin. All OTCs were explanted at PND 2. The table summarizes the firing characteristics of the neurons 6–14 DIV A B C D E DIV Firing rate (Hz) Median of ISI (ms) Mode of ISI (ms) Burst ratio Burst rate (bursts / min) 7 0.19 3369 1449 2.1 – 6 0.14 5761 51 1.2 – 14 0.18 2450 25 1.6 – 7 0.27 534 20 2.8 – 7 0.66 51 27 5.3 2.3 and neurons fire at larger intervals; in addition, the burst ratio decreases. The differences are not significant, however. The proportions of Sp-neurones remain low and fewer neurons are classified as SpB-neurones at 8–14 DIV than at 6–7 DIV. A spike train typically seen at the end of the second week in vitro is shown in Fig. 1C. 155 Fig. 2. Spontaneous activity of single neurons in OTCs of rat occipital cortex $15 DIV. (A–F) Spike time histograms (binwidth 1 s, duration 10 min). Underlined segments of A–F are shown at an enlarged time scale in A9–F9 (binwidth 40 ms, duration 25 s). The scale at the left hand side of each histogram indicates the number of action potentials / bin. All OTCs were explanted at PND 2. All neurons shown were classified as SpBneurons. The table summarizes the firing characteristics of the neurons 15–51 DIV A B C D E F DIV Firing rate (Hz) Median of ISI (ms) Mode of ISI (ms) Burst ratio Burst rate (bursts / min) 51 1.21 53 15 27.1 0.3 41 2.1 109 72 15.5 4.8 51 3.73 48 46 49.7 0.4 21 0.68 194 77 25.7 2.3 36 1.53 54 22 33.2 0.5 41 0.59 126 34 5.9 3.5 Beginning with the third week in vitro the firing intensity progressively increases and stable values are reached during the fourth week in vitro. Neurons fire more vigorously and at shorter intervals. Most neurons fire in the burst mode as indicated by the significant increases in burst ratio beyond the third week in vitro; these increases are independent from increases in firing rate. The mean firing rate doubles from the first week to stages beyond the third week although due to the large scatter in values the increase is not significant (P.0.05). Interestingly, the burst ´ , K. Albus / Developmental Brain Research 123 (2000) 151 – 164 D. Echevarrıa 156 Table 2 Developmental changes in firing intensity and temporal structure of spike trains in OTCs of rat occipital cortex a Days in vitro 6–7 8–14 15–21 22–28 29–35 .35 No. batches No. OTCs No. neurons Firing rate (Hz) 4 5 16 0.34 (0.21–1.25) 242 (125–677) 105 (36–174) 3 (1.7–5.7) 8.6 (3.2–10.8) 2 4 16 0.13 (0.09–0.23) 2904 (723–5135) 30 (23–440) 1.8 (1.6–3.4) 14.1 (9.8–24.8) 5 7 52 0.30 (0.17–0.80) 189 (74–544) 53 (21–112) 6.7 (2.6–18.3) 19.2 (10.9–34.6) 1.01 (0.70–1.72) 4 6 65 0.63 (0.24–1.28) 66(2) (23–60) 26(2) (20–48) 17.2(1) (7.2–27.9) 28.7(1) (17–49.3) 0.67 (0.46–1.08) 4 5 67 0.5 (0.24–1.67) 89 (61–164) 30(2) (23–60) 16.4(1) (9.8–70.3) 37.7(1) (27.2–56.9) 0.75 (0.44–1.13) 2 5 44 0.63 (0.26–1.77) 116 (54–189) 38 (15–51) 10.3(1) (6.1–24.2) 18.8(1) (10.5–47) 0.77 (0.52–1.76) Median of ISI (ms) Mode of ISI (ms) Burst ratio Burst ratio / firing rate Burst rate (burst / min) 0.74 (0.41–2.1) a The numbers for firing rate, median value and modal value of ISI distribution, burst ratio, burst ratio / firing rate and burst rate give the median and (in parentheses) the 25th and 75th percentile of the respective distributions. Median values from stages beyond 8 DIV significantly higher or lower (P value ,0.05) than median values of data from OTCs 6–7 DIV are marked with (1) or (2), respectively. rate does not change during time in vitro. This indicates that in spite of the increase in firing intensity the timing of bursts present shortly after explantation is maintained at later developmental stages in vitro. Further changes in firing behavior concern burst duration and the internal structure of bursts (see below). 3.2. Development of SBA in activity-deprived OTCs 3.2.1. Acute effects of elevating Mg 21 concentration The effects of elevating the Mg 21 concentration on SBA of single neurons were tested in three OTCs in which 65 neurons were recorded. Each OTC was studied for about 8 Fig. 3. Development of spontaneous activity in OTCs of rat occipital cortex. (A–D) Intra- and interbatch variability in the proportion of Sp-neurones (white bars) and NonSp-neurons (black bars) during development in vitro. Each histogram represents recordings made at different DIV in OTCs explanted from one hemisphere. All OTCs were explanted at PND 2. Numbers above the histograms indicate numbers of recorded neurons per OTC; numbers underneath histograms indicate DIV. ´ , K. Albus / Developmental Brain Research 123 (2000) 151 – 164 D. Echevarrıa h and the average number of sites explored per OTC was 26. As a control 10 randomly distributed sites were explored for SBA before changing the Mg 21 concentration. Then magnesium levels were elevated in the superfusion medium and after waiting for about 1 h in two of the three explants the effects of this manipulation on SBA were investigated by exploring at least 10 sites / OTC. Application of superfusion medium containing 5 mM or 11 mM magnesium had completely abolished SBA in these cases. After returning to normal superfusion medium the explants were again explored with microelectrodes and Sp-neurones were found in normal proportions. However, the activity of these neurons as judged on the basis of average firing rate, median value and modal value of ISI distribution and burst ratio was lower than in age-matched controls and did not recover within 2 h. In the third explant therefore, the time course of the effects of 11 mM Mg 21 were studied in more detail. About 20 randomly distributed sites were explored within 15–45 min after elevating the magnesium level. A few Sp-neurones were detected 15–30 min, and no SBA was found 30–45 min after starting with 11 mM Mg 21 . After returning to normal superfusion medium SBA recovered to about normal levels within 3 h. From these findings we concluded that 5 mM or 11 mM magnesium in the superfusion medium effectively block SBA in OTCs of rat occipital cortex. In order to reliably block SBA over longer periods OTCs were grown with 11 mM. 3.2.2. Active sites and firing variability SBA in activity-deprived OTCs was tested at recovery day zero (Rec(0)) and up to 6 days after returning to normal concentrations of Mg 21 in the growth medium (Rec(1) to Rec(6)). At Rec(0) recording was started after 3 157 h of superfusing the explant with normal Mg 21 concentrations. Methodical and experimental data for the recovery period are summarized in Table 3. The proportions of active sites / sites explored are on average 84.4613.6% (mean6S.D.). Comparisons with age-matched cases from the normal development population (22–28 DIV; Table 1) reveal slightly higher proportions of active sites in the deprived OTCs. The differences in variability of SBA over time (CV120s; Table 3) during the recovery period (Table 3) and between pooled recovery data and control cases and age-matched controls (Table 4) were not significant. 3.2.3. SBA and temporal organisation of spike trains in activity-deprived OTCs The activity blockade induces a number of abnormal firing characteristics, which distinguish deprived from normally grown explants. Throughout the recovery period activity-deprived neurons discharge sluggishly at low mean rates. The differences in mean firing rates and median interspike interval values between recovery data and data from controls (s.h.) and age-matched controls, respectively (Table 4), are significant. In contrast, modal values of ISI distributions are not different between recovery data and control cases. This assumes that activity-deprived neurons fire their rarely occurring action potentials in form of short bursts at long intervals. In fact, standardizing burst ratio with respect to firing rate reveals that the proportion of neurons discharging at shorter intervals is significantly higher in the activity-deprived cases than in the controls (Tables 3 and 4). A further persistent abnormal property of activity-deprived OTCs is a burst rate only about half of that present in the controls (Tables 3 and 4). On the other Table 3 SBA in activity-deprived OTCs of rat occipital cortex (experiment 2)a Recovery stage Rec(0) Rec(1–4) Rec(5–6) Days in vitro No. batches No. OTCs No. neurons Active sites (%; mean6S.D.) CV120s (mean6S.D.) Sp-neurons (%) SpB-neurons (%) Firing rate (Hz) Median of ISI (ms) Mode of ISI (ms) Burst ratio Burst ratio / firing rate Burst rate (burst / min) 18–27 3 9 173 80.3615 78.5644.4 33.1 100 0.12 (0.07–0.24) 383 (160–841) 28 (7–487) 12.4 (8.7–18.2) 108 (23–168) 0.45 (0.24–0.8) 20–27 3 3 72 95.465 102.4650.4 47.2 90 0.12 (0.11–0.22) 437 (120–1022) 9 (4–610) 11 (8.1–18.8) 82 (33–169) 0.35 (0.26–0.45) 28–29 2 2 40 8961.4 68.5640.5 90 86.2 0.18 (0.1–0.34) 471 (346–623) 296 (18–491) 12.3 (8.9–17.7) 84 (32–143) 0.49 (0.37–0.72) a Data from the recovery period (0 to 6 days) was grouped into Rec(0), Rec(1–4) and Rec(5–6). Average recording time (all neurons) was 10.764.1 min (mean6S.D.). For the quantitative analysis only spontaneous active units were included which were recorded longer than 4 min (Rec(0) N527, Rec(1–4) N520 and Rec(5–6) N529). The differences in activity parameters between recovery periods were not significant. For further information see footnote to Table 2. 158 ´ , K. Albus / Developmental Brain Research 123 (2000) 151 – 164 D. Echevarrıa Table 4 Effects of activity blockade on the development of SBA in OTCs of rat occipital cortex a Pooled recovery data Control s.h. Days in vitro (18–29) (20–30) Age-matched normal controls (22–35) No. neurons CV120s (mean6S.D.) Firing rate (Hz) Median of ISI (ms) Mode of ISI (ms) Burst ratio Burst ratio / firing rate SpB-neurons (N) Burst rate (burst / min) 76 81646 0.1 (0.14–0.24) 421 (316–637) 80.2 (7.5–495.1) 11.6 (8.6–18.5) 89 (48–147) 70 0.4 (0.3–0.7) 52 65636.9 0.5 (0.28–1.22) (1) 89 (61–149) (2) 30 (24–54) 16.8 (9.9–36.8) (1) 35 (23–57) (2) 59 0.7 (0.4–1.2) (1) 132 64.8636.5 0.6 (0.2–1.3) (1) 78 (44–172) (2) 27.3 (19.5–50) 14 (5.7–24.5) (1) 32 (20–50.3) (2) 113 0.7 (0.4–1.1) (1) a Data from the recovery period following activity blockade were pooled and compared with control s.h. data and data from normally grown age-matched explants. Data from control (s.h.) cases were not different from age-matched normal controls (P value .0.05). Median values from control (s.h.) cases and age-matched controls significantly higher or lower (P value ,0.05) than median values of pooled recovery data are marked with (1) and (2), respectively. Control s.h., OTCs from the same hemispheres providing the explants for the chronic activity blockade experiments. hand, the mean burst duration is about the same as measured in age-matched controls (Table 5). Other abnormalities in firing behavior in activity-deprived OTCs disappear during the recovery period (Table 3). The proportions of Sp-neurones are 33% at Rec(0) and thus resemble proportions seen during the first and second week of normal development (see Table 1). Later in the recovery period the proportions of Sp-neurones rapidly increase and reach normal levels at Rec(5–6). The high proportions of SpB-neurones recorded at Rec(0) decreased to normal levels (about 80%) at Rec(5–6). Typical examples of the temporal organization of spike trains in activity-deprived OTCs are documented in Fig. 4. In many respects the STHs compiled with a binwidth of 1 s (Fig. 4A–F) resemble the STHs seen in OTCs normally grown for 2–3 weeks. The neurones fire action potentials in the form of discrete bursts, albeit at intervals longer than in the controls. Between bursts either irregular sparse spontaneous activity (Fig. 4C, D, F) or no spontaneous activity at all (Fig. 4A, B, E) occurs. Bursts have either a tonic (Fig. 4A, D, F) or a phasic-tonic appearance (Fig. 4B, C, E) and they last from about 200 ms to 20 s; the intraburst firing rate is lower than in control cases and seldomly exceeds 10 Hz. In 90% of the spike trains with discrete burst patterns one type of burst with rather constant duration and intraburst firing rate is repeated at similar intervals. Spike trains with two or more different burst types as observed during later stages of normal development in vitro (see Fig. 2D) were not found. Striking differences between activity-deprived neurones and normally grown age-matched neurones are noted if bursts are analysed at an enlarged time scale (binwidth 40 ms; Fig. 4A9–F9). In normally developing neurones (Fig. 2A9–F9) often a burst consists of a short phasic component followed by a longer lasting tonic discharge. In activitydeprived neurones a longer ‘tonic’ burst component consists of a number of phasic subbursts usually lasting less than 1 s which are repeated at short, regular intervals (Fig. 4A9, C9, D9, F9). The firing rate during a subburst may exceed 50 Hz which is similar to firing rates measured during the initial phasic burst components in normally grown neurones (Fig. 2A9, C9). The number of neurones firing in the ‘phasic subburst’ mode decreases during recovery from activity blockade. Table 5 Firing characteristics of single neurons in activity-deprived (recovery) and normally developing OTCs of rat occipital cortex a Latency 1 peak ISI (ms) Latency 2 peak ISI (ms) Width central peak ACG (ms) Latency central peak ACG (ms) Burst duration (s) Recovery 6–14 DIV 15–28 DIV .28 DIV 12 (8–29) 380 (120–600) 200 (60–950) 20 (10–30) 8 (6–11) 28 (16–115) (1) 140 (52–340) 390 (180–1000) 30 (27.5–65) (1) 3.5 (0.95–10) 32 (20–50) (1) 96 (60–185) (2) 1000 (650–1250) (1) 50 (30–67.5) (1) 8 (5–12.5) 36 (20–48) (1) 88 (48–165) (2) 1100 (800–1400) (1) 50 (30–100) (1) 7.5 (5–13) a Normal development data were grouped into 6–14 DIV, 15–28 DIV and .28 DIV (comprising data from OTCs 29–51 DIV). The numbers give the median and (in parentheses) the 25th and 75th percentile of the distribution per stage. The latency of peaks in ISI distributions and of the central peak in ACGs correspond to the position of the bin at peak maximum. The width of the central peak in ACGs (binwidth 10 ms, time difference axis 3 s) is measured as the difference between lagtime zero and the lagtime at which the peak fell below the expected value for a homogeneous distribution. The burst duration is determined only for SpB-neurons; it is measured either directly from the spike time histograms (binwidth 1s) or from ACGs with a binwidth of 50 ms or 100 ms (time difference axis 15 s and 30 s, respectively). The number of neurons per group is .50 with exception of group 6–14 DIV, in which 22 ISI distributions and eight SpB-neurons were evaluated. Median values of normal development data significantly higher or lower (P value 0.05) than median values of recovery data are marked with (1) and (2), respectively. Firing characteristics did not change during normal development with the exception of the width of the central ACG peak which increased significantly from 6–14 DIV to .28 DIV. ´ , K. Albus / Developmental Brain Research 123 (2000) 151 – 164 D. Echevarrıa Fig. 4. SBA in activity-deprived OTCs of rat occipital cortex. (A–F) Spike time histograms (binwidth 1 s, duration 10 min). Underlined segments of A–F are shown at an enlarged time scale in A9–F9 (binwidth 40 ms, duration 25 s). The scale at the left hand side of each histogram indicates the number of action potentials / bin. All OTCs were explanted at PND 2. All neurons shown were classified as SpB-neurons. The table summarizes the firing characteristics of the neurons Recovery A B C D E F DIV Recovery day Firing rate (Hz) Median of ISI (ms) Mode of ISI (ms) Burst ratio Burst rate (bursts / min) 28 4 0.26 691 640 13.8 0.7 24 0 0.17 332 10 8.4 0.6 20 0 0.21 69 6 15.3 0.2 25 2 0.19 62 6 5.4 0.2 29 6 0.41 165 18 22.8 1.1 28 5 1.83 33 7 12.3 0.6 3.3. Temporal characteristics of spike trains: a comparison between activity-deprived and normally grown OTCs The temporal organisation of spike trains was further investigated by analysing interspike interval distributions (ISIs; Fig. 5) and autocorrelogram histograms (ACGs; Fig. 6) in more detail. The developmental increase in firing intensity of normally grown neurones (see Table 2) is reflected in the ISI 159 distributions (Fig. 5; 6–14 DIV, 15–51 DIV). At 6–14 DIV many neurones often display multiple peaks in their ISIs, thus firing at intervals shorter than 100 ms and in addition, at intervals longer than 200 ms. At later developmental stages the proportions of longer intervals and accordingly, the number of secondary and tertiary peaks in the ISI distributions are reduced; in addition the spikes sampled over a 10-min period have increased in numbers (Fig. 5, 15–51 DIV). ISI distributions in activity-deprived OTCs, in particular from recovery days 0–3 (Fig. 5, Recovery) have a short latency peak followed after at least 200 ms by distinct second and third peaks; these distributions resemble distributions recorded during the first and second week in vitro. However, in activity-deprived OTCs the latency of the first peak is shorter and that of the second peak longer than in normally developing OTCs (Table 5). This type of distribution reflects the ‘phasic subburst’ firing mode described above (Fig. 49, C9, D9, F9). ACG histograms in the activity-deprived cases resemble that seen in normally developing OTCs 6–14 DIV in normal development (Fig. 6). In particular, proportions of ACGs without a significant peak (Fig. 6, 6–14 DIV, A; see Section 2) are similar between both groups (35% and 38%, respectively); in the normal development cases longer than 14 DIV these proportions account to less than 15%. ACGs with significant peaks which in addition to a tonic, slowly decaying flank display a distinct initial phasic component (Fig. 6: Recovery (E, F); 6–14 DIV (D, E); 15–51 DIV (C, E, F)), account to 39% in both activity-deprived cases and normal developing cases 6–14 DIV. At later stages of normal development these proportions exceed 59%. Tonic ACGs with a slowly decaying flank lacking a distinct initial phasic component (Fig. 6: Recovery (A), 15–51 DIV (A, B, D)) increase in proportions during normal development (from 20 to 28%) and are rare in the activitydeprived cases (10%) in which they are seen only at recovery days 4–6. Proportions of ACGs (Fig. 6: Recovery (B–D), 6–14 DIV (B, C)) dominated by a short initial phasic component preferentially occur during early phases of recovery from deprivation and during 6–14 DIV of normal development. The quantitative evaluation of ACG data (Table 5) supplements the conclusions made from analysing ISI distributions. The width of the central ACG peak is smaller and the latency of the central ACG peak shorter in activitydeprived neurones than in normally developing neurones. The differences most likely reflect the differences in the temporal organisation of bursts: many activity-deprived neurones discharge in the phasic ‘subburst’ mode, whereas normally grown neurones prefer phasic-tonic or tonic burst discharges. We occasionally observed neurones oscillating at stable frequencies (see Fig. 6; Recovery (F), 15–51 DIV (E)) ranging between 0.7 and 50 Hz. The proportions of these neurones were 7.5% in the activity-deprived cases, and between 4.4 and 5.9% in the normal development cases. At 160 ´ , K. Albus / Developmental Brain Research 123 (2000) 151 – 164 D. Echevarrıa Fig. 5. Interspike interval (ISI) distributions of spike trains in activity-deprived (recovery) and normally grown (6–14 DIV, 15–51 DIV) OTCs of rat occipital cortex. The ISI distributions relate to the spike trains shown in Fig. 4A–F (recovery), Fig. 1A–E (6–14 DIV) and Fig. 2A–F (15–51 DIV). The scale on the left hand side of each histogram gives the number of intervals per bin; binwidth is 10 ms. 6–14 DIV the two oscillating neurones found had a frequency of 1.1 Hz and 0.7 Hz, respectively. 4. Discussion 4.1. Spontaneous activity and firing rates The activity types we have defined in OTCs of the rat occipital cortex have also been recorded by others in organotypic [11,16,18,19] and explant cultures from rat neocortex [10,17,20,27] and rat hippocampus [36,55], in primary dissociated cell cultures from rat neocortex cells [28,42,44] and rat and mouse hypothalamic neurons [34,35]. In explant cultures from parietal cortex 3–6 weeks in vitro the proportion of cells discharging spontaneous action potentials (20%) was much lower than the proportions (.80%) recorded in our OTCs at that developmental stage. The difference might be due to differences in the methods for preparation and maintenance of OTCs; pieces of tissue were 1 mm thick and culturing and recording was performed under equilibration with 5% CO 2 [54]. The developmental increase in the proportion of spontaneously active neurons we have seen in the neocortical OTCs is a common phenomenon found in vivo [4,30] and in vitro [18,20,34,35,42]. The slight developmental decrease in the proportion of bursting neurons reported in primary dissociated cell cultures of neocortex [28] has not been confirmed by others [44]. The moderate increase in mean firing rates with time in vitro is in agreement with another report [11]; according to our data the increase is not significant, however. Similar findings were reported from primary dissociated cell cultures of neocortex [42,44] and hypothalamus [35]. The firing rates of spontaneously active neurons in our OTCs are similar to firing rates reported from the isolated neocortex of the turtle [8] and neocortical primary dissociated cell cultures [42,44]. SBA seems to develop to higher levels in vivo than in vitro. In the occipital cortex of ´ , K. Albus / Developmental Brain Research 123 (2000) 151 – 164 D. Echevarrıa 161 Fig. 6. Autocorrelograms (ACGs) of spike trains in activity-deprived (recovery) and normally grown (6–14 DIV, 15–51 DIV) OTCs of rat occipital cortex. The ACGs relate to the spike trains shown in Fig. 4A–F (recovery), Fig. 1A–E (6–14 DIV) and Fig. 2A–F (15–51 DIV). The scale on the left hand side of each histogram gives the number of action potentials in one bin; binwidth is 10 ms. The two bars at the end of each ACG give the expected value for a homogeneous distribution, and the chance level (P value50.001), respectively. the unanesthetized kitten firing rates reach up to 3 Hz during the second postnatal week and about 8 Hz in the adult. In the freely moving rat firing rates increase to 3 Hz at postnatal days 11 and 12 [37] and up to 10 Hz in the adult [3]. The delay in the development of SBA during the second week in vitro could be caused by an imbalance in the growth and differentiation of inhibitory and excitatory systems and / or cell death of particular neuronal populations [9,32,41]. Habets et al. [28] have presented evidence that the developmental increase in firing variability in primary dissociated cell cultures might be due to an arrest in the maturation of inhibitory interneurones. However firing variability remained constant during development in vitro in our OTCs. We have found that with time in vitro neurons tend to organize their discharges in the form of discrete bursts. In addition, the proportions of SpB-neurones as well as the intensity of the bursts, i.e. the intraburst frequency, increase. The developmental patterns of SBA in OTCs of neocortex are therefore, similar to that in OTCs of hippocampus where a progressive time-dependent onset of spontaneous long lasting group discharges was found resembling interictal activity in vivo [36,55]. Hippocampal OTCs older than DIV 30 contained high proportions (60– 100%) of neurons with seizure-like activity. Seizure-like activity is suggested to develop due to alterations in synaptic efficiency, neuronal conductances and / or changes in intrinsic properties of neurons [36]. Interestingly, blockade of inhibition does not seem to be a prerequisite of seizure-like activity [27]. The spontaneous synchronous discharges in neocortical OTCs seem to 162 ´ , K. Albus / Developmental Brain Research 123 (2000) 151 – 164 D. Echevarrıa reflect the combined efforts of many neurons distributed all over the explant; different neurons — although not necessarily connected synaptically — participate in the population phenomenon underlying depolarization shifts [16,27]. 4.2. Spontaneous activity as an epigenetic factor in the development of bioelectric activity in vitro We have shown that the main effect of activity blockade is a halt or delay in the development of SBA. Interestingly, in primary dissociated cell cultures of neocortex [42] activity blockade increases mean firing rates, intraburst firing frequencies and burst rates, and decreases burst duration. Thus, in these cultures the predominant effect of an activity blockade during development in vitro is a facilitation of SBA and not a depression as in OTCs of rat occipital cortex. The difference most probably reflects the differences in normal development of SBA between both preparations. In primary dissociated cell cultures the burst ratio and interval dependencies decrease during development in vitro [28,42]. Spontaneous action potentials mostly in the form of discrete bursts but at rather long interspike intervals are observed towards the end of the first week in vitro and stereotyped burst patterns are present only during the second week in vitro [28]. In our OTCs higher burst ratios and stereotyped bursts became dominant only at later developmental stages. Also, the analysis of autocorrelograms indicated an increase rather than a decrease in the occurrence of interval dependencies during development in vitro. Since the discharge patterns in activity-deprived dissociated cell cultures closely resemble firing patterns seen after acute administration of picrotoxin it was suggested that GABA-ergic synaptic inhibition is ineffective in cortical networks grown under conditions which prevent SBA [14,42,45]. The finding in OTCs of rat neocortex that activity blockade strongly reduces the expression of GAD mRNA supports this idea [46]. Surprisingly, the impairment of the GABA-system did not result in an increase in firing intensity; in contrast the authors reported significantly reduced average firing rates shortly after the blockade of SBA had been terminated, a finding which is in accordance with our results. The halt or delay in the development of SBA in activitydeprived OTCs may be caused therefore, by a retardation of the structural and functional development of excitatory systems. This suggestion is supported by anatomical studies. A significant reduction in the number of dendritic spines and thus, excitatory synaptic connections in pyramidal neurons has been reported in neocortical OTCs silenced by applying both PTX and TTX [1]. The neuronal outgrowth and the synapse formation of synaptic ultrastructure is retarded in dissociated cell cultures from neocortex treated with TTX [52]; retardation of axonal outgrowth was induced by TTX in the goldfish optic nerve [23]. As a consequence of such changes the increase in network density during development in vitro would be delayed. The lack of longer lasting tonic burst discharges in most activity-deprived neurons which is present during the early recovery period indicates a decrease in the number and efficiency of both, excitatory inputs and a synaptic NMDA receptor-mediated component. In normally developing OTCs this component has been shown to contribute to the maintenance and amplification of seizure-like events [27,36,48]. A possible reason for the depression of neuronal activity induced by chronic exposure to magnesium (Mg 21 ) might be cell death affecting both excitatory and inhibitory neurons [7]. However, according to the findings of others [41] high levels of Mg 21 prevent spontaneous neuronal degeneration in the hippocampus; widespread neuronal degeneration occurs only after removal of the blocking agent. One would expect that a substantial loss of neurons would reduce the number of neurons detected when randomly probing the explant with a microelectrode [6]. However, in our activity-deprived OTCs the density of viable neurons as judged by the proportions of active sites was around 85% and the same as in normally grown controls. In addition, the proportions of SpB-neurones and Sp-neurones had returned to normal values after 5 to 6 days of recovery. The activity depression seen in our activity-deprived OTCs is therefore, unlikely to be caused by cell death. Acknowledgements We wish to thank Susanne Lausmann for technical assistance and Erik Aarnoutse and Dr. Gras for providing software for data evaluation. This work was supported by the Alexander von Humboldt Foundation (grant SPK0025), the Gobierno Vasco, Spain (grant no. BIF94.046) and the BMBF (FKZ 0310954). References [1] C.M. Annis, D.K. O’Dowd, R.T. Robertson, Activity-dependent regulation of dendritic spine density on cortical pyramid neurons in organotypic slice cultures, J. Neurobiol. 25 (1994) 1483–1493. [2] A. Antonini, M.P. Stryker, Development of individual geniculocortical arbors in cat striate cortex and effects of binocular impulse blockade, J. Neurosci. 13 (1993) 3549–3573. [3] M. Armstrong-James, The functional status and columnar organization of single cells responding to cutaneous stimulation in neonatal rat somatosensory cortex S1, J. Physiol. 246 (1975) 501–538. [4] M. Armstrong-James, K. Fox, The physiology of developing cortical neurons, in: E.G. Jones, A. Peters (Eds.), Development and Maturation of Cerebral Cortex. Cerebral Cortex, Vol. 7, Plenum Press, New York, 1988, pp. 237–266. [5] R.E. Baker, M.A. Corner, A.M.M. Habets, Effects of chronic suppression of bioelectric activity on the development of sensory ´ , K. Albus / Developmental Brain Research 123 (2000) 151 – 164 D. Echevarrıa [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] ganglion evoked responses in spinal cord explants, J. Neurosci. 5 (1984) 1187–1192. R.E. Baker, M.A. Corner, T. Lammertse, E. Furth, Some functional effects of suppressing bioelectric activity in fetal mouse spinal cord-dorsal root ganglion explants, Exp. Neurol. 94 (1986) 426– 430. R.E. Baker, J.M. Ruijter, Chronic blockade of bioelectric activity in neonatal rat neocortex in vitro: physiological effects, Int. J. Dev. Neurosci. 9 (1991) 321–329. M.G. Blanton, A.R. Kriegstein, Spontaneous action potential activity and synaptic currents in the embryonic turtle cerebral cortex, J. Neurosci. 11 (1991) 3907–3923. M. Caesar, T. Bonhoeffer, J. Bolz, Cellular organization and development of slice cultures from rat visual cortex, Exp. Brain Res. 77 (1989) 34–244. M.-C. Calvet, Patterns of spontaneous electrical activity in tissue cultures of mammalian cerebral cortex vs. cerebellum, Brain Res. 69 (1974) 281–295. S. Cardoso de Oliveira, A. Reissner, O. Klostermann, P. Wahle, Development of spontaneous activity in organotypic cultures, in: ¨ 22nd Gottingen Neurobiology Meeting, Vol. 2, 1994, p. 537. S.M. Catalano, C.J. Shatz, Activity dependent cortical target selection by thalamic axons, Science 281 (1998) 559–562. M.A. Corner, S.M. Crain, Patterns of spontaneous bioelectric activity during maturation in culture of fetal rodent medulla and spinal cord tissues, J. Neurobiol. 3 (1972) 25–45. M.A. Corner, G.J.A. Ramakers, Spontaneous firing as an epigenetic factor in brain development. Physiological consequences of chronic TTX and PTX exposure on cultured rat neocortex neurons, Dev. Brain Res. 65 (1992) 57–64. I. Costero, C.M. Pomerat, Cultivation of neurons from the adult human cerebral and cerebellar cortex, Am. J. Anat. 89 (1951) 405–467. S.M. Crain, Development of ‘organotypic’ bioelectric activities in central nervous tissues during maturation in culture, Int. Rev. Neurobiol. 9 (1966) 1–43. S.M. Crain, M.B. Bornstein, E.R. Peterson, Maturation of cultured embryonic CNS tissues during chronic exposure to agents which prevent bioelectric activity, Brain Res. 8 (1968) 363–372. S.M. Crain, M.B. Bornstein, Organotypic bioelectric activity in cultured reaggregates of dissociated rodent brain cells, Science 176 (1972) 182–184. S.M. Crain, M.B. Bornstein, Early onset of imhibitory functions during synaptogenesis in fetal mouse brain cultures, Brain Res. 68 (1974) 351–357. S.M. Crain, Neurophysiologic Studies in Tissue Cultures, Raven Press, New York, 1976. J. Del Castillo, L. Engbaek, The nature of the neuromuscular block produced by magnesium, J. Physiol. 124 (1954) 370–384. ´ K. Albus, Development of spontaneous bioelectric D. Echevarrıa, activity in organotypic cultures of rat occipital cortex with and without spontaneously occurring action potentials, Soc. Neurosci. Abs. 24 (2) (1998) 1306. D.L. Edwards, B. Grafstein, Intraocular tetrodotoxin in goldfish hinders optic nerve generation, Brain Res. 269 (1983) 1–14. ¨ B.H. Gahwiler, A.M. Mamoon, C.A. Tobias, Spontaneous bioelectric activity in cultured cerebellar purkinje cells during exposure to agents which prevent synaptic transmission, Brain Res. 53 (1973) 79. ¨ B.H. Gahwiler, Organotypic monolayer cultures of nervous tissue, J. Neurosci. Methods 4 (1981) 329–342. Q. Gu, Y. Liu, R.H. Dyck, V. Booth, M.S. Cynader, Effects of tetrodotoxin treatment in LGN on neuromodulatory receptor expression in developing visual cortex, Dev. Brain Res. 106 (1998) 93–99. M.J. Gutnick, B. Wolfson, F.J. Baldino, Synchronized neuronal activities in neocortical explants cultures, Exp. Brain Res. 76 (1989) 131–140. 163 [28] A.M.M.C. Habets, A.M.J. Van Dongen, F. Van Huizen, M.A. Corner, Spontaneous neuronal firing patterns in fetal cortical networks during development in vitro: a quantitative analysis, Exp. Brain Res. 69 (1987) 43–52. [29] M.J. Hogue, Human fetal brain cells in tissue culture: their identification and motility, J. Exp. Zool. 106 (1947) 85–107. [30] P.R. Huttenlocher, Development of cortical neuronal activity in neonatal cat, Exp. Neurol. 17 (1967) 247–262. [31] R.R. Kalil, M.W. Dubin, G. Scott, L.A. Stark, Elimination of action potentials blocks the structural development of retinogeniculate synapses, Nature 323 (1986) 156–158. [32] R. Kapoor, C.B. Jaeger, R. Llinas, Electrophysiology of the mammalian cerebellar cortex in organ culture, Neuroscience 26 (1988) 493–507. [33] T. Kasamatsu, R. Adey, Immediate effects of total visual deafferentation on single unit activity in the visual cortex of freely behaving cats. I. Tonic excitability changes, Exp. Brain Res. 20 (1974) 157–170. [34] P. Legendre, A. Tixier-Vidal, J.L. Brigant, J.D. Vicent, Electrophysiology and ultrastructure of mouse hypothalamic neurons in culture: a correlative analysis during development, Dev. Brain Res. 43 (1988) 273–285. [35] D.S.F. Ling, R.E. Petroski, W. Chou, M. Geller, Development of spontaneous electrical activity by rat hypothalamic neurons in dissociated culture, Dev. Brain Res. 53 (1990) 276–282. [36] C.J. McBain, P. Boden, R.G. Hill, The rat hippocampal slices in vitro display spontaneous epileptiform activity following long-term organotypic cultures, J. Neurosci. Methods 27 (1989) 35–49. [37] M. Mirmiran, M. Corner, Neuronal discharge patterns in the occipital cortex of developing rats during active and quiet sleep, Dev. Brain Res. 3 (1982) 37–48. [38] T.H. Murphy, L.A. Blatter, W.G. Wier, J.M. Baraban, Spontaneous synchronous synaptic calcium transients in cultured cortical neurons, J. Neurosci. 12 (1992) 4834–4845. ¨ [39] G. Paxinos, I. Tork, L.H. Tecott, K.L. Valentino, Atlas of the Developing Rat Brain, Academic Press, San Diego, 1991. [40] A.A. Penn, P.A. Riquelme, M.B. Feller, C.J. Shatz, Competition in retinogeniculate patterning driven by spontaneous activity, Science 279 (1998) 2108–2112. [41] L.D. Pozzo Miller, N.K. Mahanty, J.A. Connor, D.M. Landis, Spontaneous pyramidal cell death in organotypic slice cultures from rat hippocampus is prevented by glutamate receptor antagonists, Neuroscience 63 (1994) 471–487. [42] G.J.A. Ramakers, M.A. Corner, A.M.M.C. Habets, Development in the absence of spontaneous bioelectric activity results in increased stereotyped burst firing in cultures of dissociated cerebral cortex, Exp. Brain Res. 79 (1990) 157–166. [43] G.J.A. Ramakers, G.J. Boer, Chronic suppression of bioelectric activity and cell survival in primary cultures of rat cerebral cortex: biochemical observations, Eur. J. Neurosci. 3 (1991) 154–161. [44] G.J.A. Ramakers, C. de Wit, P.S. Wolters, M.A. Corner, A developmental decrease in NMDA-mediated spontaneous firing in cultured rat cerebral cortex, Int. J. Dev. Neurosci. 11 (1993) 25–32. [45] G.J.A. Ramakers, H. Van Galen, M.G. Feenstra, M.A. Corner, G.J. Boer, Activity-dependent plasticity of inhibitory and excitatory amino acid transmitter systems in cultured rat cerebral cortex, Int. J. Dev. Neurosci. 12 (1994) 611–621. [46] A. Reissner, S. Cardoso de Oliveira, O. Klostermann, P. Wahle, Blockade of spontaneous activity in rat occipital cortex cultures ¨ downregulates GAD mRNA expression, in: 23rd Gottingen Neurobiology Meeting, Vol. 2, 1995, p. 102. [47] R.V. Riccio, M.A. Matthews, The postnatal development of the rat primary visual cortex during optic nerve impulse blockade by intraocular tetrodotoxin: a quantitative electron microscopic analysis, Dev. Brain Res. 20 (1985) 55–68. [48] M.M. Segal, Epileptiform activity in microcultures containing one excitatory hippocampal neuron, J. Neurophysiol. 65 (1991) 761– 770. 164 ´ , K. Albus / Developmental Brain Research 123 (2000) 151 – 164 D. Echevarrıa [49] F.J. Seil, R. Drake-Baumann, Reduced cortical inhibitory synaptogenesis in organotypic cerebellar cultures developing in the absence of neuronal activity, J. Comp. Neurol. 342 (1994) 366–377. [50] M.P. Stryker, W.A. Harris, Binocular impulse blockade prevents the formation of ocular dominance columns in cat visual cortex, J. Neurosci. 6 (1986) 2117–2133. [51] I. Thompson, C. Holt, Effects of intraocular tetrodotoxin on the development of the retinocollicular pathway in the Syrian hamster, J. Comp. Neurol. 282 (1989) 371–388. [52] F. Van Huizen, H.J. Romijn, A.M.M.C. Habets, Synaptogenesis in rat cerebral cortex is affected during chronic blockade of spontaneous bioelectric activity by TTX, Dev. Brain Res. 19 (1985) 67–80. [53] J.M. Woijtowicz, K.C. Marshall, W.J. Hendelman, Depression by magnesium ion of neuronal excitability in tissue culture of central nervous system, Can. J. Physiol. Pharmacol. 55 (1977) 367–372. [54] B. Wolfson, M.J. Gutnick, F.J. Baldino, Electrophysiological characteristics of neurons in neocortical cultures, Exp. Brain Res. 76 (1989) 122–130. [55] B. Zipser, S.M. Crain, M.B. Bornstein, Directly evoked paroxymal depolarizations of mouse hippocampal neurons in synaptically organized explants in longterm cultures, Brain Res. 60 (1973) 489–495.