Skip to main content
NCBI home page
The Journal of Physiology logoLink to The Journal of Physiology
. 2004 May 21;558(Pt 1):193–211. doi: 10.1113/jphysiol.2004.061416

Burst generation in rat pyramidal neurones by regenerative potentials elicited in a restricted part of the basilar dendritic tree

Bogdan A Milojkovic 1, Mihailo S Radojicic 1, Patricia S Goldman-Rakic The Late 1, Srdjan D Antic 1
PMCID: PMC1664906  PMID: 15155788

Abstract

The common preconception about central nervous system neurones is that thousands of small postsynaptic potentials sum across the entire dendritic tree to generate substantial firing rates, previously observed in in vivo experiments. We present evidence that local inputs confined to a single basal dendrite can profoundly influence the neuronal output of layer V pyramidal neurones in the rat prefrontal cortical slices. In our experiments, brief glutamatergic stimulation delivered in a restricted part of the basilar dendritic tree invariably produced sustained plateau depolarizations of the cell body, accompanied by bursts of action potentials. Because of their small diameters, basolateral dendrites are not routinely accessible for glass electrode measurements, and very little is known about their electrical properties and their role in information processing. Voltage-sensitive dye recordings were used to follow membrane potential transients in distal segments of basal branches during sub- and suprathreshold glutamate and synaptic stimulations. Recordings were obtained simultaneously from multiple dendrites and multiple points along individual dendrites, thus showing in a direct way how regenerative potentials initiate at the postsynaptic site and propagate decrementally toward the cell body. The glutamate-evoked dendritic plateau depolarizations described here are likely to occur in conjunction with strong excitatory drive during so-called ‘UP states’, previously observed in in vivo recordings from mammalian cortices.

Several lines of evidence suggest that basal dendrites play a very important role in cortical information processing. Studies that combine physiological and histological techniques have shown that connections between layer V large pyramidal cells are mediated by synaptic junctions placed mainly on the basal dendrites (Deuchars et al. 1994; Markram et al. 1997). Local excitatory contacts are essential in a process known as ‘recurrent excitation’, which is thought to be the cellular substrate of persistent neuronal activity (Goldman-Rakic, 1995; Compte et al. 2000; Goldman et al. 2003). Basal and proximal oblique dendrites are ideally suited to participate in recurrent excitation in that they comprise approximately two-thirds of the total membrane area of a neurone. Based on dendritic spine counts, it has been estimated that within the boundaries of layer V, basal and oblique branches receive approximately 65% of the total number of excitatory synaptic contacts (Larkman, 1991). Interestingly, Lucifer yellow injections showed that those cortical pyramidal neurones, which exhibit UP and DOWN states in vivo, have noticeably rich basal dendritic arbors (Steriade et al. 1993a). In vitro models of cortical rhythmic recurrent activity unequivocally showed that synaptic inputs arriving on basal and proximal oblique dendrites within the boundaries of layers V and VI are the major source of excitatory drive during the UP state. After the amputation of the apical dendritic tree, neurones void of apical synaptic inputs (in brain slices containing only layers V and VI) reliably generated UP and DOWN states (Sanchez-Vives & McCormick, 2000). Finally, the biophysical properties of basal dendrites (relatively short and directly attached to the soma–axon decision-making point) render them important subcellular specializations in cortical networking (Archie & Mel, 2000; Poirazi & Mel, 2001; London et al. 2002).

Recently, two groups (Schiller et al. 2000; Oakley et al. 2001a) have shown that strong activation of postsynaptic glutamate receptors by glutamate uncaging/iontophoresis can generate dendritic spikes in basal dendrites. These studies addressed dendritic integration using somatic recordings and calcium imaging. Due to the complex geometry of basal dendrites and unknown distribution of voltage-gated membrane conductances, somatic electrical recordings can neither detect nor explain the details of the non-linear synaptic processing in remote dendritic segments (Steriade et al. 1993a). Calcium imaging, on the other hand, has the excellent spatial resolution to tackle dendritic function (Wei et al. 2001). However, calcium concentration in the dendritic cytosol is a complex function of several variables including: the membrane potential (Ross et al. 1987; Lasser-Ross et al. 1991), the state of calcium-permeable glutamate receptors (MacDermott et al. 1986), and the second messenger pathways that govern the release of calcium from intracellular stores (Llano et al. 1991; Finch & Augustine, 1998; Nakamura et al. 1999). Finally, the slower dynamics of calcium signals (Markram et al. 1995) do not allow detection of the more rapid repolarizations of membrane potential transients that are a key feature of dendritic sodium action potentials (Stuart & Sakmann, 1994) and dendritic calcium spikes (Markram & Sakmann, 1994; Schiller et al. 1997). Because of multiple sources and slow dynamics, calcium imaging is not ideally suited for studying the integration of synaptic inputs when they produce isolated dendritic potentials, and even less suitable when synaptic inputs coincide with somatic bursts. Given the normal background-firing rate of prefrontal cortical pyramidal neurones in vivo (Funahashi et al. 1989; Williams & Goldman-Rakic, 1995; Miller et al. 1996), synaptic integration in distal dendrites rarely occurs in the absence of somatic action potentials (APs). When synaptic potentials and back-propagating action potentials interact in distal dendrites, it is useful to turn from calcium- to voltage-sensitive dye imaging (Zecevic, 1996).

Somatic depolarizations obtained in pioneering experiments on basal dendrites were subthreshold for initiation of somatic action potentials (Schiller et al. 2000; Oakley et al. 2001a,b). In the present study, however, brief glutamate iontophoresis directed on basal dendrites regularly produced long-lasting somatic depolarizations accompanied by bursts of action potentials. These glutamate-evoked events resembled in vivo UP states (Cowan et al. 1994; Branchereau et al. 1996; Lewis & O'Donnell, 2000) in three important details: (1) the amplitude of the somatic plateau depolarization; (2) the duration of the plateau phase; and (3) the number and frequency of APs per plateau event. Being the major recipients of intralaminar (recurrent) excitatory connections (Gilbert, 1983; Cauller & Kulics, 1991; Kritzer & Goldman-Rakic, 1995), it is likely that during in vivo UP states basal dendrites of layer V pyramidal cells often process suprathreshold excitatory inputs, and thereby the dendritic plateau potentials described here are likely to occur in conjunction with UP states.

Methods

Brain slices

Sprague-Dawley rats (P21–42) were anaesthetized with halothane and decapitated according to an animal protocol approved by Yale University Animal Care and Use Committee. Coronal slices (300 μm) were cut (NVSL, Campden Instruments) from the frontal lobe in ice-cold solution, which contained (mm): 125 NaCl, 26 NaHCO3, 10 glucose, 2.3 KCl, 1.26 KH2PO4, 2 CaCl2 and 1 MgSO4 (pH 7.4 when bubbled with 95% O2–5% CO2). After an initial incubation of 45 min at 35°C, slices were stored in a holding chamber at 20–21°C. All experiments were performed at 29–34°C using two heaters: (1) on the bottom of the recording chamber, and (2) in the in-flow pipe (dual temperature controller 344B Warner Instrument Corp.).

Electrophysiology

A fixed stage microscope (Zeiss Axioskop 2FS) with a 250 W xenon arc lamp and two CCD cameras mounted to the microscope body were used in all experiments. Physiological recordings were made on the medial edge of the brain slice (medial prefrontal cortex) from visually identified layer V pyramidal cells (infrared video microscopy, Dage-MTI IR-1000). To ensure that these cells had intact apical dendrites, we flipped slices so that the apical trunk descended into the tissue. Neurones were patched using a Narishige micromanipulator MMN-21 and 7 MΩ pipettes pulled from borosilicate glass capillaries (1.5 mm outer diameter, 0.86 mm inner diameter, Warner Instrument Corp.) on a Sutter p-97 electrode puller. Current clamp recordings were made using a Multiclamp 700A (Axon Instruments) and digitized at 1 or 5 kHz with two AD boards: (1) NeuroPlex (RedShirtImaging), and (2) Digidata 1322A (Axon Instruments).

Rhodamine tracing

For tracing of the basal dendrites deep in the slice (Fig. 1; see also supplementary Fig. S1 available online only), we used tetramethylrhodamine dextran 3000. Patch pipettes were loaded with rhodamine (0.1 mm) dissolved in a filtered intracellular solution containing electrolyte concentrations (mm): 135 potassium gluconate, 10 Hepes, 2 MgCl2, 3 Na-ATP, 0.3 Na-GTP, 10 creatinine phosphate (pH 7.3 adjusted with KOH; osmolarity 275 mosmol kg−1, adjusted with H2O). Approximately 25 min after the whole-cell break-through, basal dendrites were visible under green excitation light. To prevent the phototoxicity of intracellularly applied rhodamine, the illumination episodes were kept to a minimum. This was typically 3–5 s for each position of the stimulating electrode.

Figure 1. Glutamate-evoked bursts of action potentials.

Figure 1

Open in a new tab

A, composite microphotograph of a layer V pyramidal neurone filled with rhodamine. Arrowheads indicate three different positions of the glutamate pipette: (1) 90 μm (2) 125 μm, and (3) 45 μm from the cell body. B, whole-cell somatic recordings of glutamate-evoked responses of the cell shown in A, at stimulus positions 1 (top), 2 (middle), and 3 (bottom). In this and following figures ‘glut.’ marks time points of iontophoretic glutamate pulses. Stimulus parameters (intensity 1.8 μA, duration 5 ms, and frequency 1 Hz) were the same in stimulus positions 1–3. ‘c.i.’, somatic current injection (200 pA, 250 ms). Action potentials in positions 2 and 3 are truncated. Resting membrane potential Vm = −63 mV. C, the amplitudes, durations and action potential counts during the second (II) and third (III) glutamate-evoked plateaus were normalized to the first (I) plateau (n = 24). Means and standard deviations are presented in the form of a bar diagram. D, a portion of the sweep shown in B (position 1) is displayed here on a fast time base: I, II and III mark the first, second, and third glutamate-evoked events. Note that after the termination of the action potential burst, the cell body remains in a sustained depolarized state for several hundred milliseconds.

Voltage-sensitive dye imaging

For voltage imaging we used positively charged styryl molecules JPW3028 and JPW1114, designed by J. P. Wuskell and L. Loew (University Connecticut, Farmington, CT, USA).

Voltage-sensitive dye JPW3028 (0.8–1.5 mm), or JPW1114 (0.8 mm), was dissolved in regular intracellular solution (see above, omitting rhodamine). To prevent leakage of the lipophilic fluorescent dyes out of the patch pipette and contamination of brain tissue around the neurone, the tips of patch pipettes were front-filled with dye-free intracellular solution. This procedure greatly reduces the background fluorescence (Antic et al. 1999), as well as the actual concentration of dyes in the cytoplasm. Optical measurements of dendritic membrane transients were carried out with an 80 × 80 pixel cooled CCD camera (NeuroCCD, RedShirtImaging) operated under the NeuroPlex software (written in IDL, Research Systems Inc., Boulder, CO, USA), which was also used to drive the mechanically isolated uncased shutter (35 mm Uniblitz) and the stimulator. JPW dye signals (excitation 520 ± 45 nm, dichroic 570 nm, and emission 610 nm) were sampled at 2.7 and at 1 kHz. In optical recordings of single action potentials (Fig. 4), temporal signal averaging was carried out using a spike-triggered mode; but all sweeps were saved and inspected for consistency. Off-line data analysis (spatial averaging and digital filtering) was performed using Gaussian (low-pass) and Butterworth (high-pass) filters. Unless specified differently, 6–16 neighbouring pixels were selected from each region of interest for spatial averaging. The amplitude of the optical signal (ΔF/F) was calculated as ‘average change in light intensity/average resting light intensity’. Due to signal saturation in the centre of the soma, for the somatic region of interest (ROI), pixels were selected from the periphery of the soma. In some neurones, after the experiment, nine sweeps were averaged in the absence of the stimuli and subtracted from the physiological records to correct the bleaching of the dye. In order to reduce the phototoxicity of the voltage-sensitive dyes (Antic et al. 1999) the following steps were taken: (1) after loading of the cell body (∼30 min), the dye-loading pipette was withdrawn from the slice (outside-out patch); (2) the injected dye was allowed approximately 2 h to diffuse from the injection site (soma) into the dendritic tree; (3) during the dye injection and incubation period, the microscope and overhead lights were extinguished; (4) prior to voltage-sensitive measurements, neurones were re-patched with dye-free solution; (5) focusing of the stained neurones was done with 2–5% neutral density optical filters inserted into the excitation illumination path; (6) the duration of the optical recording sweeps was limited to 90 ms in those experiments, in which temporal averaging was used to improve signal-to-noise ratio (Fig. 4B). In other experiments excitation illumination was limited to 1 s and intersweep intervals were kept above 30 s.

Figure 4. Decremental propagation of dendritic spikes.

Figure 4

Open in a new tab

Same recordings as in Figs 3C and D. This time the Gaussian filter was set to a 600 Hz cut-off to alleviate the effects of low-pass filtering on the amplitude of the action potential-associated optical signals. A, a single sweep recording of glutamate-evoked APs at resting membrane potential Vm = −56 mV. B, while keeping the neurone in the same position and focus, a somatic AP was triggered with direct current injection (30 ms, 200 pA). Each trace is a spike-triggered average of 4 sweeps. C, a single-sweep recording of glutamate-evoked dendritic plateau potential in an artificially hyperpolarized neurone (Vm=−67 mV). In the distal dendrite (ROIs 6 and 5), the amplitude of the dendritic plateau is similar to the amplitude of the back-propagating action potential, while in the proximal dendritic region (ROIs 2 and 3) the dendritic potential is a small fraction of the AP. The number in parenthesis indicates the distance from the cell body. White lines superimposed on traces 5 and 6 are the same signals filtered at 50 Hz cut-off to determine the amplitude of the slow component (dashed lines).

Glutamate and synaptic stimulation

Sharp glass pipettes (40 ± 10 MΩ) were filled with 200 mm sodium glutamate (pH 9), attached to the motorized micromanipulator (Sutter Instruments P-285), and positioned at less than 15 μm from the distal dendritic membrane, using fluorescent and infrared microscopy. An Iso-Flex (A.M.P.I) stimulus isolation unit, controlled by a Master 8 programmable pulse-generator, was used to deliver (5 ms duration, and 0.8–3.0 μA amplitude) negative constant current pulses. A minus sign is omitted in the text.

In experiments where the release of endogenous glutamate was triggered by shocking synaptic preterminals, sharp electrodes were replaced with 7 MΩ patch pipettes filled with extracellular solution. After insuring that the electrode tip was positioned at less than 25 μm from the distal dendrite, electric shocks were delivered with fixed duration (100–200 μs), while current amplitude varied in the range of 10–90 μA.

Data analysis

The somatic plateau amplitude was measured in AxoScope8.1 (Axon Instruments) as the difference between the depolarization peak following the afterhyperpolarization of the last AP in the plateau event and the baseline. The duration of the somatic plateau depolarization was measured at half-amplitude (half-width). Averaged data are presented in the text as means ± standard deviation. Statistical significance was determined by Student's paired t test (P < 0.01), unless otherwise specified.

Results

Glutamate-evoked plateau depolarizations

Our findings were obtained by stimulating layer V pyramidal neurones with brief glutamate pulses (5 ms) delivered to a circumscribed segment of basal dendrites. Fluorescent dye (rhodamine) was applied intracellularly to facilitate positioning of the glutamate iontophoretic electrode (Fig. 1A). The neurones exhibited little spontaneous activity in vitro and remained quiescent throughout the recording episodes (∼45 min) in 2 mm extracellular calcium (resting membrane potential Vm= 62.2 ± 5.1 mV, n = 30). However, glutamate iontophoretic ejections (1–3 μA, 5 ms) delivered on distal dendritic segments (60–100 μm from the cell body; 79.8 ± 9.3 μm, (mean ± s.d.), n = 28) produced characteristic plateau potentials, on top of which action potentials rode (Fig. 1B, and supplementary Fig. S1B, position 1). The amplitude of the somatic plateau depolarization was in the range of 12–23 mV (17.05 ± 3.69 mV).

In order to assess the robustness and variability of dendritic responses, glutamate pulses were delivered on basal dendrites at a frequency of 1 Hz (Fig. 1D). The mean absolute (and normalized) plateau amplitudes of the first, second, and third events in the train were 17.16 ± 3.72 mV (100%), 16.94 ± 3.73 mV (98.67 ± 4.77%), and 17.05 ± 3.75 mV (99.47 ± 5.33%, n = 24), respectively. The mean absolute (and normalized) half-widths of the first, second and third plateau phases were 361.5 ± 92.3 ms (100%), 380.6 ± 105.9 ms (104 ± 3.4%), and 393.8 ± 105.3 ms (108.7 ± 4.25%), respectively. The mean numbers of spikes per plateau for the first, second and third glutamate iontophoresis stimulations were 4.39 ± 1.4, 4.07 ± 1.4, and 3.96 ± 1.3, respectively (Fig. 1C). No statistical difference was detected in plateau amplitude between the first and second (Student's paired t test, PI-II= 0.1051), nor between the first and third (PI-III= 0.5040) glutamate-evoked event in a 1 Hz train. The durations of the second and third plateau depolarizations were consistently longer than the duration of the first plateau in the train. Although relatively small on average (4.76 ± 3.45% and 8.71 ± 4.25%), these differences were statistically significant (Student's paired t test, PI-II < 0.00001, PI-III < 0.00001, n = 24). The data presented so far indicate that glutamate iontophoresis delivered on basal dendrites can produce a characteristic somatic signal made up of a slow (plateau depolarization) and a fast (action potentials) component, and that this neuronal response is very robust at a 1 Hz stimulation frequency.

In nine neurones the glutamate stimulation site was 100–145 μm away from the soma. In this group of neurones, glutamate-evoked somatic plateau depolarizations (11.7 ± 2.4 mV, n = 9) were not accompanied by sodium spikes. In order to test whether the amplitude of somatic depolarizations actually depended on the distance of the dendritic stimulation site from the cell body, in one group of neurones (n = 6 neurones, n = 10 dendrites) identical glutamate pulses (intensity, duration, and frequency) were delivered at two locations along the same dendritic branch, using the same iontophoretic pipette and stimulus parameters. The mean distances of the proximal and distal stimulation site from the centre of the soma were 78.2 ± 11.8 and 120.3 ± 14.7 μm, respectively. In all neurones tested in this way, glutamate pulses resulted in a brisk plateau depolarization of the cell body, regardless of the distance from the stimulation site. However, distal stimulation sites (Fig. 1B, position 2) were invariably less successful (10 out of 10 dendrites) in bringing neurones to the action potential firing threshold than the proximal stimulation sites (position 1). The mean plateau amplitude, plateau duration, and spike count per plateau for the proximal and distal stimulation sites along the same basal dendrite were, respectively, 18.16 ± 3.3 mV, 335.4 ± 57.8 ms and 3.9 ± 1.52 spikes for the proximal site and 12.22 ± 4.9 mV, 346.9 ± 101.0 ms and 0.1 ± 0.32 spikes for the distal site. Student's paired t test analysis showed that the difference in plateau amplitude was statistically significant (Pprox-dist= 0.0007), while the change in plateau duration was not (Pprox-dist= 0.7002). These data suggest that local glutamate stimulation delivered at two locations on the same dendrite evokes dendritic plateau depolarizations of similar duration, and presumably similar amplitude. Due to their different electrotonic distances from the soma, the two dendritic signals undergo a different degree of attenuation. In consequence, the cable-filtered amplitude of the distally evoked signal is smaller than the amplitude of the signal evoked at the proximal stimulation site (Fig. 1B and supplementary Fig. S1B).

One important concern is that glutamate in our experiments diffused from the ejection site and directly affected the somatic membrane. In this scenario the observed difference in signal amplitude between proximal and distal stimulation protocols (Fig. 1 and supplementary Fig. S1) simply reflects the concentration gradient of the ejected glutamate. To test this hypothesis, the tip of the glutamate pipette was moved from the dendritic stimulation site to an area closer to the cell body but void of dendrites. We deliberately sought a wide gap between dendritic branches. In 3/3 neurones tested in this way, previously suprathreshold glutamate pulses (Fig. 1B, position 1), produced miniature postsynaptic signals at 45–55 μm from the soma (position 3). These results showed that the glutamate did not diffuse from the ejection site (positions 1 and 2) to act directly on the cell body. Instead, the observed somatic plateau depolarizations (positions 1 and 2) were the consequence of dendritic stimulation.

Voltage-sensitive dye imaging of glutamate-evoked dendritic potentials

In order to study the actual dendritic membrane potential transients, instead of rhodamine, we applied voltage-sensitive dyes intracellularly (Fig. 2A). Optical signals were recorded with an 80 × 80 pixel camera (Antic, 2003). During glutamate-evoked somatic bursts, the dendrites that were targeted by the glutamate pipette (target dendrites) exhibited a fast-rising plateau depolarization, on top of which back-propagating APs rode (Fig. 2C, region of interest ROI 2). Each somatic sodium spike was represented by a corresponding peak on the dendritic ‘crest’ potential (Fig. 2A inset, arrows). At glutamate ejection sites (ROI 2) the amplitude of the slow component (plateau) was a significant fraction of the AP-associated optical signal, such that the mean plateau/AP amplitude ratio was 0.65 ± 0.11 (n = 12). Non-target dendrites (ROIs 3 and 4) also exhibited a slow component (plateau depolarization) in conjunction with back-propagating spikes, but the plateau/AP ratio in non-target dendrites (0.30 ± 0.08) was significantly different (P < 0.00001, n = 12) from the plateau/AP ratio measured in target dendrites, or in the cell body (0.21 ± 0.07). These data suggest that distal segments of target dendrites were the cellular compartment where plateau depolarizations were generated, while slow potentials recorded in non-target basal branches were cable-filtered somatic plateaus, i.e. the slow component of the electrical signal spread centrifugally from the cell body into inactive basal dendrites.

Figure 2. Optical monitoring of glutamate-evoked bursts.

Figure 2

Open in a new tab

A, composite microphotograph of a layer V pyramidal neurone filled with the voltage-sensitive dye JPW1114. Upper inset: optical signal obtained at the stimulation site filtered temporarily with a low-pass Gaussian filter with a 100 Hz cut-off. Arrows mark peaks of back-propagating action potentials. Lower inset: optical signal from the target dendrite has two components: a slow component (plateau) and fast component (action potentials). B, the area of fluorescent image where optical signals were sampled as indicated by a rectangle in A. The illumination aperture was partially closed (circle in A). A schematic drawing marks the position of the glutamate iontophoresis pipette 105 μm from the soma. C, a single glutamate pulse (5 ms, 2.2 μA) was delivered onto the distal dendritic segment, and membrane potential changes were recorded electrically from the cell body (ROI 1) and optically (ROIs 2–4) from basal dendrites, according to the map in B. Optical traces 2, 3, and 4 are a spatial average of 29, 31, and 23 pixel outputs from neighbouring detectors in the region of interest marked by boxes 2-4, respectively. The proximal and distal edges of boxes 2,3 and 4 are 90-140, 60-110, and 55-90 μm from the centre of the soma, respectively. Signals were sampled at 1 KHz frame-raqte with no filtering.

To characterize the initiation and propagation of dendritic events in the absence of back-propagating spikes, we imaged basal dendrites while the cell body was artificially hyperpolarized to prevent the initiation of somatic APs (n = 5). In hyperpolarized neurones, a glutamate pulse evoked a square-shaped dendritic plateau potential with a fast onset, long-lasting plateau phase (100–500 ms) and an abrupt collapse at the end of the plateau phase (Fig. 3D). In two neurones, an initial spikelet was detected at the beginning of the plateau (Fig. 3D, ROI 6). In all neurones (5 out of 5), the amplitude (ΔF/F) of the slow component of the glutamate-evoked optical signal in the target dendrites gradually faded in the somatopetal direction (Fig. 3C and D, ROIs 3 and 2). The average cable-filtered plateau measured in the cell body of hypopolarized neurones was 15.7 ± 3.4 mV (n = 5).

Figure 3. Dendritic plateau potentials.

Figure 3

Open in a new tab

A, composite microphotograph of a layer V pyramidal neurone filled with the voltage-sensitive dye JPW3028. Schematic drawing marks the position of the glutamate pipette 165 μm from the centre of the soma. B, the area of fluorescent image captured with a low-resolution (80 × 80 pixels) data acquisition camera, indicated by the rectangle in A. C, single (5 ms) glutamate pulse (glut.) was delivered in the distal dendritic segment and membrane potential changes were recorded electrically from the cell body (ROI 1) and optically (ROIs 2–9) along the basal dendrites according to the map in B. D, same as in C except the cell body was hyperpolarized with DC current to prevent initiation of somatic action potentials, and glutamate iontophoretic current was increased from 1.0 to 1.2 μA. Each optical trace is a spatial average of 9 pixel outputs from neighbouring detectors in the region of interest (ROI). Signals were sampled at a 2.7 kHz frame-rate and digitally filtered off-line with a low-pass Gaussian filter with a 200 Hz cut-off. The intensity of glutamate current and somatic resting membrane potential are shown on the top of the panel.

Decremental forward propagation of glutamate-evoked dendritic plateau potentials

Intracellular voltage-sensitive dyes cannot be used to determine the absolute amplitude (in mV) of the electrical transients in distal dendritic segments (Antic et al. 1999). However, voltage-sensitive dyes can be used to detect a relative amplitude change between signals obtained from the same dendritic segment in consecutive recording trials.

Since the concentration and partition of the voltage-sensitive dye in a given dendritic segment is unlikely to change in 1–2 min, the sensitivity of voltage-sensitive dye measurements obtained from the same dendritic segment (using the same detector pixel) remains constant between subsequent recordings. Any difference in the optical signal amplitude between two subsequent sweeps is therefore due to the difference in the amplitude of the membrane potential transient.

While keeping the fluorescent neurone in a fixed position and plane of focus between recording trials, we found that the amplitude of the glutamate-evoked plateau potential was 0.67 ± 0.10 (n = 5) of the amplitude of the back-propagating AP (Fig. 4). Computer simulation based on realistic neuronal morphology and multisite voltage-sensitive dye measurements predicted that in basal dendritic segments 150 μm from the soma, the amplitude of the back-propagating action potential (AP) is at least 60 mV (Antic, 2003). Thus, the amplitude of the glutamate-evoked plateau at 150 μm from the soma (Fig. 4C, ROI 5) was estimated to be at least 40 mV.

At the initiation site (ROIs 6 and 5), glutamate-evoked plateaus (Fig. 4A and C) had rather similar amplitudes as back-propagating action potentials (Fig. 4B). In proximal dendritic segments (ROIs 3 and 2), however, the glutamate-evoked event was several times smaller than the back-propagating action potential. A significant amplitude difference between glutamate-evoked plateaus and action potential signals was also observed in non-target dendrites (ROIs 7, 8, and 9). The simplest interpretation of these findings is that glutamate-evoked plateau potentials, triggered in a distal region of a target dendrite, propagate decrementally towards the soma, and from there they spread into non-target basal branches.

Glutamate threshold for initiation of dendritic spikes

Voltage-sensitive dye imaging was used to monitor dendritic membrane potentials in response to a gradual increase in the amount of ejected glutamate. During smaller iontophoretic currents, the glutamate-evoked change in dendritic membrane potential grew gradually with iontophoretic current (Fig. 5C, trials 1–3). In all neurones tested (n = 9), increasing the intensity of the glutamate iontophoretic current above some ‘critical’ value evoked square-shaped potentials, characterized by a fast onset and a plateau phase, on top of which there were one or two back-propagating APs (Fig. 5C, trial 4). Thus, a small change in glutamate current produced a ‘jump’ (spike) in the electrical response of the distal dendritic segment (Fig. 5D and E). The average half-width of the just-threshold dendritic plateau (Fig. 5E, arrow) was 106 ± 42 ms (n = 9). An increase in iontophoretic current above the critical glutamate concentration (threshold) increased the duration of the plateau potential (n = 9/9), but did not contribute significantly to its amplitude (Fig. 5C, trials 5–9, and Fig. 5E).

Figure 5. A sudden jump in the glutamate-evoked dendritic potential.

Figure 5

Open in a new tab

A, composite microphotograph of a layer V pyramidal cell stained with JPW1114. B, the recording field for voltage-imaging measurements is marked by a white rectangle in A. Schematic drawing indicates the position of the glutamate-filled glass pipette, 135 μm from the soma. The neurones shown in this and the following figure (Fig. 6) were not re-patched after incubation with the voltage-sensitive dye. C, the intensity of glutamate current was increased gradually and voltage-sensitive dye signals were sampled from the distal dendritic segment in 12 trials. The inter-trial interval was 35 s. Nine out of 12 trials were selected for display. The intensity of glutamate iontophoretic current is displayed above each trace. Optical signals (sweeps 1–9) are products of spatial averaging (48 pixels), Gaussian low-pass (50 Hz) and Butterworth high-pass (0.4 Hz) digital filtering. Pixels for spatial averaging were selected from the dendritic region starting at 105 and ending at 165 μm from the cell body (marked by white polygonal line in B). There was no temporal averaging. D, two consecutive measurements (sweeps 3 and 4 of panel C) are superimposed and expanded to show a sudden jump in signal amplitude. Asterisks in this and the previous panel indicate somatic action potentials, detected in voltage-sensitive dye recordings from the cell body (not shown). E, amplitudes of glutamate-evoked dendritic optical signals are plotted versus gradually increasing intensities of glutamate iontophoretic current. Both the amplitudes and glutamate current intensities are normalized with respect to the point on the graph showing a sudden jump (arrow). At some ‘critical glutamate intensity’ the amplitude of the dendritic optical signal nearly doubled (arrow) compared to previous stimulus intensity. The mean value of the ‘critical glutamate intensity’ was 1.18 ± 0.22 μA (n = 9). The mean amplitude of the dendritic signal (ΔF/F) obtained at critical glutamate intensity was 1.96 ± 0.42%(n = 9). Inline graphic, pooled data from 9 neurones. O, data points from the experiment presented in C.

Synaptically evoked dendritic plateau potentials

To test whether the glutamate concentrations used in experiments described so far were compatible with endogenous sources, i.e. glutamate stored in presynaptic axon terminals; we applied single electric shocks in the vicinity of distal basal dendrites (15–25 μm from the dendritic shaft). In 6/6 neurones synaptic stimulation evoked a sustained depolarization at the stimulation site. The average half-width of the synaptically evoked dendritic depolarization was 152 ± 63 ms (n = 6). In 3/6 neurones (stimulated with a single shock), one or two back-propagating APs rode on the top of the plateau (Fig. 6B, ROI 5). Optical signals from the neighbouring (non-target) dendrites (Fig. 6C, ROIs 8–10) also captured a plateau depolarization with back-propagating APs superimposed (asterisks). However, the onset of the plateau depolarization in a non-target dendrite was considerably slower than the synaptically evoked plateau in the target branch (Fig. 6C), thus suggesting that the synaptic depolarization generated in the target dendrite spread over to the non-target branch via the somatic compartment (cable-filtered electrical signal).

Figure 6. Spatio-temporal profile of membrane potential in the dendritic tree during synaptically evoked bursts of action potentials.

Figure 6

Open in a new tab

A, composite microphotograph of the basal dendritic tree stained with JPW3028. Schematic drawings indicate the position of the extracellular stimulation electrode and selection of pixels for display in B. The distance of the stimulation electrode from the centre of the soma is ∼150 μm. B, optical signals collected along primary (ROIs 2, 3 and 4) and secondary (ROIs 5, 6 and 7) branches of the target dendrite are aligned with signals from the apical (ROI 1) and non-target basal dendrite (ROI 8). The timing of the extracellular pulse (0.2 ms, 60 μA) is marked on the stimulus line (stim.). C, optical signal from the target dendrite (ROI 5) was digitally filtered (Gaussian low-pass, 50 Hz cut-off) and superimposed with signals sampled along the non-target basal branch (ROIs 8, 9 and 10) to show the difference in dynamics between the synaptically evoked dendritic potential (ROI 5) and the cable-filtered dendritic depolarization (ROIs 8, 9 and 10). Asterisks indicate the timing of the somatic APs.

In 3/6 neurones, a single synaptic stimulus did not trigger somatic action potentials. In the example shown in Fig. 7, an electric shock, delivered at a distance of 15 μm from the basal dendrite, 170 μm distal to the soma (Fig. 7B, double arrow) triggered a plateau potential in the target dendrite (Fig. 7A, trial 1, ROIs 6–10), but not in the dendritic branch located 105 μm lateral to the stimulation site (ROIs 11–13). The somatic whole-cell electrode (ROI 1) detected a small initial spikelet (5 mV amplitude), and a slow excitatory (EPSP) component, which overlapped with a fast inhibitory postsynaptic potential (IPSP). The synaptically evoked dendritic potential (trial 1, ROI 9) was characterized by (i) a fast initial spikelet, and (ii) a plateau phase of 110 ms, and an abrupt collapse (breakdown) at the end of the plateau phase. The dendritic spike propagated decrementally toward the soma (ROIs 5 and 4), resulting in a complete loss of signal in the most proximal dendritic segment (ROI 3).

Figure 7. Synaptically evoked dendritic plateau potentials – single shock.

Figure 7

Open in a new tab

A, trial 1: simultaneous electrical (ROI 1) and optical (ROIs 2–13) recordings of membrane potential upon a single extracellular shock applied 15 μm from the dendrite, 170 μm from the soma. Control trials 2 and 3: somatic current injection produced an action potential, which invaded both basal dendrites. Trial 4: same stimulus as in trial 1. Trial 5: stimulus intensity was increased from 50 to 60 μA. Arrow indicates a depression in the optical signal (hypopolarization), which coincides with the fast IPSP detected in the soma (ROI 1). Signals are sampled at 1 kHz and digitally filtered offline (Gaussian low-pass, 200 Hz). Each optical trace is a spatial average of 9–12 pixel outputs, from neighbouring pixels selected in the region of interest. B, a compound microphotograph of the neurone taken after the experiment. The rectangle indicates the area in the object field from which fast optical measurements were acquired. C, a low-resolution image of the part of the dendritic tree where optical signals were acquired. The double arrow, in this and the previous panel, indicates the position of the stimulating extracellular electrode. Single arrows indicate dendritic segments (regions of interest) where pixel outputs were spatially averaged to produce the optical traces displayed in A. A movie presentation of these data is available online (Supplementary Material).

Next, with the neurone in focus, we initiated action potentials by injecting current directly into the soma (trials 2 and 3). These control trials established that both dendritic branches were electrically responsive, and that the lack of signal in the most proximal region of the target dendrite and along the non-target dendrite during synaptic stimulation was not due to impaired excitability. Second, these trials provided a comparison between the amplitude of the back-propagating action potential and the synaptically evoked dendritic plateau potential (plateau/AP = 0.65 ± 0.16, n = 3).

At this point we have to address two important concerns regarding the measurements of synaptically evoked membrane potential transients. First, dendritic potentials evoked by single synaptic shock (Fig. 6) seem less squared off than the glutamate-evoked responses observed with glutamate stimulation (Fig. 3). Second, single-shock synaptic stimulations seem incapable of producing the stable long-lasting somatic plateau potentials observed upon glutamate iontophoresis (Fig. 1 and supplementary Fig. S1). The differences between iontophoretically evoked responses and actual synaptic responses could be significant. For example, iontophoretic applications of glutamate were likely to cause much more prolonged elevations in glutamate concentration than those evoked by synaptic transmission. Such prolonged elevation might perhaps have recruited different membrane mechanisms (e.g. different species of voltage-gated conductances) than the rapid depolarization that would be expected to result from activation of synapses. To allay these concerns, in the next series of experiments we used trains of synaptic stimuli instead of single shocks (Fig. 8). In addition, we intermittently applied glutamate iontophoresis and synaptic stimulation on the same segment of a basal dendrite (Figs 8 and 9). Optical measurements (Fig. 8B, grey trace) show that a characteristic square-shaped dendritic plateau potential can be evoked with synaptic stimulation only. This result was obtained in 8 out of 14 neurones tested. The duration of the somatic plateau depolarization recorded by whole-cell pipettes was in the range 140–590 ms (average duration 312 ± 173 ms, n = 8). In these neurones, the somatic membrane was experiencing stable long-lasting depolarization for several hundred milliseconds after the last synaptic stimulus in the train (Fig. 8C, arrow). The somatic depolarization plateau collapsed, on average, 251 ± 175 ms after the last stimulus artefact.

Figure 8. Synaptically evoked dendritic plateau potentials – multiple shocks.

Figure 8

Open in a new tab

A, microphotograph of a pyramidal neurone stained with JPW1114. Schematic drawings indicate the positions of the extracellular stimulation electrode (syn.) and glutamate iontophoresis pipette (glut.). The distance of each stimulation electrodes from the centre of the soma was ∼100 μm. B, 18 neighbouring pixels were selected from area marked by a white rectangle in A, and spatially averaged to compare a dendritic signal in response to synaptic stimulation (grey) to that evoked by glutamate (black). C, whole-cell recordings during synaptic stimulation (train of 3, frequency 50 Hz, intensity 30 μA) with a resting membrane potential of Vm = −58 mV. The arrow marks long-lasting somatic depolarization. D, whole-cell recordings during glutamate iontophoresis (duration 5 ms, intensity 1.7 μA) with a resting membrane potential of Vm = − 57 mV.

Figure 9. Comparison between synaptically and glutamate-evoked dendritic potentials.

Figure 9

Open in a new tab

A, microphotograph of a pyramidal neurone stained with JPW1114. Schematic drawings indicate the positions of the extracellular stimulation electrode (syn.) and glutamate iontophoresis pipette (glut.). The distance of each stimulation electrodes from the centre of the soma was ∼110 μm. B, voltage-sensitive dye signals from the target dendrite (ROI 1) and non-target basal branch (ROI 2) are aligned with the somatic whole-cell recording (ROI 3). The neurone was first stimulated synaptically (train of 5, frequency 50 Hz, intensity 12 μA), and then iontophoretically (single pulse, duration 5 ms, intensity 1.3 μA). The collapse of the synaptically evoked somatic depolarization (arrow) occurred 55 ms after the last shock in the train. Fast inhibitory postsynaptic potentials produced indentations on the top of the somatic plateau depolarization (3) but not on the dendritic voltage-sensitive dye recordings (1 and 2). Resting membrane potential Vm = −61 mV. C, amplitudes of dendritic voltage-sensitive dye signals (ΔF/F) upon synaptic (grey) and glutamate (black) stimulation obtained in 7 neurones are presented in the form of a bar graph. Cells 6 and 7 did not generate sodium action potentials. Each signal is a product of 15–24 spatially averaged pixels from the target basal branch. Although the number of selected pixels varied among neurones, the same group of pixels was used to compare dendritic response to synaptic and glutamate stimulation within a neurone.

In seven neurones we applied glutamate iontophoretically at the same dendritic site, which was the target of synaptic stimulation in a preceding trial. The intensity of the glutamate pulse was adjusted to match the half-width of the synaptically evoked somatic plateau depolarization (Figs 8D and 9B3). The direct comparison of glutamate-evoked and synaptically evoked optical signals in the target dendrite (Fig. 9C) showed that amplitudes of these two types of dendritic response were in relatively good agreement (Student's paired t test, P = 0.5812). The average ratio of glutamate- and synaptically evoked signals obtained from the same dendritic segment in two consecutive trials (glut/syn) was 0.98 ± 0.25 (n = 7). Five out of seven neurones used in this analysis fired three to six APs per plateau event (Fig. 8C, and Fig. 9C, cells 1–5). Since back-propagating APs are known to boost glutamate-evoked dendritic calcium in neocortical pyramidal neurones (Koester & Sakmann, 1998; Schiller et al. 1998) and neostriatal medium spiny neurones (Kerr & Plenz, 2004), a concern is raised that synaptic and iontophoretically evoked voltage signals in the present study might initially have been very different, but after being boosted by back-propagating APs these signals achieved the same amplitude (Fig. 8B). In two neurones both synaptic and glutamate stimulation failed to trigger somatic APs. This allowed us to measure the amplitudes of the glutamate-evoked and synaptically evoked optical signals, which were not contaminated with back-propagating APs (Fig. 9B). In the absence of back-propagating sodium spikes two types of glutamatergic stimulation (synaptic and iontophoretic) produced similar membrane potential transients in basal dendrites (Fig. 9C, cells 6–7).

Discussion

Cortical UP states provide strong glutamatergic drive

In vivo intracellular recordings show that cortical pyramidal neurones and striatal medium spiny cells alternate between hyperpolarized (DOWN) and depolarized (UP) states (Wilson & Groves, 1981; Steriade et al. 1993a; Wilson & Kawaguchi, 1996). In the slices maintained in vitro, the membrane potentials of the cells correspond to that of the DOWN state, and no depolarizing (UP) episodes are seen, unless the excitability of neurones is altered by reducing the concentration of the extracellular calcium (Sanchez-Vives & McCormick, 2000; Shu et al. 2003). The UP states are based solely on patterned synaptic excitation (Steriade et al. 1993b; Cowan & Wilson, 1994; Wilson & Kawaguchi, 1996; Lampl et al. 1999; Shu et al. 2003). Here we present an experimental design in which ‘patterned synaptic excitation’ was delivered on a single basal branch of a layer V pyramidal neurone. We are asking what would be the consequence of spatial grouping of glutamatergic afferents on a basal branch (Archie & Mel, 2000). Massive synchronous synaptic excitation impinging on one part of the basal dendritic tree was here mimicked by brief (5 ms) glutamate pulse via a sharp glass pipette or by 50 Hz synaptic stimulation. Our major finding is that a single primary basal dendrite is capable of delivering enough current flow to keep the cell body in a long-lasting depolarized state (Figs 1, 2, 8, 9 and 10, and supplementary Fig. S1). Most importantly, during these glutamate-evoked depolarizations the cell body often fired bursts of APs, thus resembling the in vivo ‘UP state’ (Steriade et al. 1993b; Branchereau et al. 1996; Lewis & O'Donnell, 2000) in several important parameters: (1) plateau amplitude (12–23 mV), (2) plateau duration (200–700 ms), (3) number of APs per plateau event (2–12, median = 4), and (4) irregular firing pattern during the plateau phase (supplementary Fig. S1C, red arrows).

Figure 10. Evaluation of neuronal response to glutamate in the presence of internally applied voltage-sensitive dyes.

Figure 10

Open in a new tab

A, composite microphotograph of a layer V pyramidal neurone filled with JPW1114 (0.8 mm in the back of the pipette). Schematic drawing marks the position of the glutamate-filled pipette on the basal dendrite, 75 μm from the cell body. B, glutamate was ejected (duration 5 ms, intensity 1.8 μA) in regular intervals of 35 s. While keeping the tip of the glutamate pipette in a fixed position, the neurone was subjected to repeated optical recordings (each optical recording epoch 1.2 s). Somatic whole-cell recordings were acquired before the first illumination epoch (trial 0) and during each optical recording sweep (trials 1–11). Resting membrane potential varied ±1.1 mV around −58 mV. C, an area marked by a white rectangle in A is expanded to show the details of the basal dendritic tree where optical measurements were performed. The illumination aperture was partially closed to block strong fluorescence light from the cell body. D, 15 neighbouring pixels selected from the area marked by white rectangle in C are spatially averaged, temporarily filtered (Gaussian filter, 25 Hz cut-off), and displayed on the same time scale as the whole-cell recording shown in B. Note that repetitive exposures to excitation light (520 ± 45 nm) did not compromise dendritic response to exogenous glutamate (trials 1–11). ‘Glut.’ marks the timing of the glutamate pulse.

Voltage-sensitive dye measurements showed that the wave of depolarization always started in the distal region of the target dendrite at the stimulation site (Figs 2, 3, 4, 6 and 7). Using calcium imaging, two groups have recently concluded that basal dendrites of layer V pyramidal neurones can trigger regenerative potentials in response to glutamate stimulation (Schiller et al. 2000; Oakley et al. 2001b). Since their methods precluded the direct recording of dendritic membrane potential, Schiller et al. (2000) attempted to study dendritic transients in computer simulations. Their description of dendritic glutamate-evoked potentials is based on a number of assumptions incorporated in a set of mathematical equations (model). Our description of glutamate-evoked dendritic potentials is based on direct experimental recordings of dendritic membrane potential changes during subthreshold and suprathreshold excitatory stimulations. Voltage-sensitive dye measurements, performed simultaneously in the target and non-target basal dendrites, revealed the exact time course of the glutamate- and synaptically evoked dendritic potentials, as they initiate on the postsynaptic membrane and propagate into the cell body and neighbouring dendritic branches. Dendritic plateau potentials were studied both in isolation (Figs 3 and 9) and in conjugation with somatic APs (Figs 2–8). When somatic AP initiation was blocked by hypopolarizing current, dendritic plateau potentials showed a fast onset with or without an initial spikelet, a long-lasting plateau phase, and an abrupt collapse at the end of the plateau phase (Fig. 3D). The absolute amplitude of dendritic plateau depolarization could not have been determined with voltage-sensitive dyes, but relative measurements indicated that at 150 μm away from the cell body the amplitude of the dendritic plateau potential is approximately 2/3 of the amplitude of the back-propagating sodium spike (Figs 4 and 7).

Endogenous glutamate (glutamate stored in vesicles of presynaptic axon terminals) seems to be perfectly capable of triggering dendritic plateau potentials (Fig. 7), especially if synaptic stimulation was delivered in trains of three to five excitations (Fig. 8). Based on the good agreement in amplitude of the dendritic membrane potential transient (Fig. 9), both sources of glutamate, synaptic release or iontophoresis, engage the same postsynaptic membrane mechanism, characterized by both a hard threshold and a saturation of the dendritic depolarization at suprathreshold glutamate concentrations (Fig. 5D).

Phototoxicity

Voltage-sensitive dyes are notorious for causing physiological changes in cells (Antic et al. 1999). Is it possible that synaptically evoked and glutamate-evoked dendritic plateau potentials presented here are a consequence of a dendritic pathology caused by the detrimental effects of free radicals on the excitable membrane (Feix & Kalyanaraman, 1991; Krieg, 1992)? Three lines of evidence show that the voltage-sensitive dyes used here did not significantly alter the neuronal response to glutamatergic stimulation. First, and most importantly, glutamate-evoked somatic plateau depolarizations in rhodamine-stained neurones were not significantly different than those obtained from neurones injected with voltage-sensitive dyes, in terms of the amplitude and duration (Table 1). Second, the intensity of the glutamate iontophoretic current used in the rhodamine-filled neurones to trigger 200–300 ms long plateau potentials accompanied by two or three sodium action potentials (Ig= 1.64 ± 0.44 μA, n = 17) was not statistically different from that measured in neurones injected with voltage-sensitive dyes (Ig= 1.89 ± 0.40 μA, n = 13). Third, the glutamate-evoked neuronal response was fairly stable across successive exposures to excitation light. In the course of experiments, accumulation of free radicals during successive optical recording epochs is expected to show a larger extent of photodynamic damage in later trials. In eight neurones (filled with the voltage-sensitive dye) we kept the glutamate pipette in a fixed position and monitored the amplitude and duration of the glutamate-evoked somatic plateau depolarization (Fig. 10) across consecutive optical recording epochs (1.2 s each). Group data from these measurements show that within ∼10 s of total exposure time (9 trials), the neuronal response to glutamate was stable with respect to the amplitude and duration of the somatic plateau depolarization (supplementary Fig. S2A and B). Action potential count, on the other hand, showed a small but consistent decline with the length of exposure to strong excitation light (supplementary Fig. S2C). Taken together, the effects on the number of APs in the burst and the finding that glutamate-evoked plateau depolarization declined in both amplitude and duration if a neurone was exposed to more than 20 s of strong excitation light (not shown), indicate that the phototoxicity of currently available fluorescent probes is still a major limiting factor in the wider application of fast multisite voltage-sensitive dye recordings.

Table 1.

Comparison of glutamate-evoked plateau parameters in rhodamine- and JPW1114-filled neurones

Plateau amplitude (mV) P* Plateau duration (ms) P Glutamate current intensity (μA) P n
Rhodamine 17.9 ± 3.9 268 ± 63 1.64 ± 0.44 17
0.5689 0.6607 0.1319
JPW1114 18.7 ± 4.0 280 ± 78 1.89 ± 0.40 13
Open in a new tab
*

P values derived by Student's t test.

Basal dendrites as independent computational modules

That dendritically initiated spikes may fail to invade the cell body has been previously shown in apical dendrites of neocortical neurones (Schiller et al. 1997; Helmchen et al. 1999) and hippocampal pyramidal neurones (Golding & Spruston, 1998). Line scan calcium imaging has recently been used to monitor glutamate-evoked calcium signals in one basal dendrite at a time (Schiller et al. 2000; Oakley et al. 2001a). A full understanding of the compartmentalization of fine (small diameter) dendrites has had to await spatially well-resolved optical measurements taken simultaneously from several neighbouring branches in the visual field (Wei et al. 2001; Euler et al. 2002). In the present study, simultaneous voltage-sensitive dye recordings from several neighbouring dendritic branches provided the first documented evidence that dendritic spikes can be truly isolated electrical events in the basilar dendritic tree, i.e. confined to a single basal dendrite. In the experiment depicted in Fig. 7, a fast IPSP could have been responsible for the somatic failure to generate an AP. Appropriately targeted perisomatic inhibition has been previously suggested as a possible factor for the failure of the dendritic spike to invade the soma (Golding & Spruston, 1998). Though inhibition may be a significant factor in the example shown in Fig. 7, we also observed isolated dendritic spikes in the absence of IPSPs and/or artificial hypopolarization, during glutamate or synaptic stimulations (Figs 1B and 9B), indicating that other factors could contribute to spike propagation failures. Impedance mismatch (Luscher et al. 1994; Rapp et al. 1996) is thought to be a critical factor in spike propagation failure from the apical dendrite (a region of higher impedance) to the soma (a region of low-impedance). In the case of thin basal dendrites, the impedance mismatch is even greater (Poirazi & Mel, 2001), thereby decreasing the likelihood of dendritic spikes invading the soma and reaching the threshold for sodium AP firing (supplementary Fig. S1B).

Voltage-sensitive dye recordings performed along dendrites exposed to glutamate ejected from a pipette (Fig. 3) or glutamate released from synaptic terminals (Fig. 7) unequivocally showed that dendritic plateau spikes were relatively large at the stimulation site (2/3 of AP amplitude in the distal dendritic segment) but decay quickly on the way to the soma (Fig. 4). One important hypothesis is that voltage-gated glutamate receptor–channel currents (NMDA) contribute actively and significantly to the magnitude of the dendritic plateau spike (Schiller et al. 2000). Since glutamate is supplied very locally in distal regions, glutamate receptors in the proximal segments of the target dendrite were not activated, and therefore did not actively support dendritic potentials. The lack of active support combined with large impedance mismatch caused these potentials to decay rapidly in the somatopetal direction (Fig. 4), and provide the cell body with just 15–20 mV of sustained depolarization (supplementary Fig. S1B, blue arrow).

Spatial and temporal segregation of synaptic inputs

Our results suggest that basal dendrites of prefrontal pyramidal neurones are endowed with a membrane mechanism that can turn circumscribed synaptic activity into a long-lasting somatic plateau potential (supplementary Fig. S1C, blue arrows). Synaptic terminals spatially clustered in the same dendritic segment may use such a mechanism to drive neuronal output by a coherent synchronous discharge. Similar dendritic subunits have been previously suggested (by Shepherd, 1972; Koch et al. 1989; Finch & Augustine, 1998; Takechi et al. 1998; Schiller et al. 2000; Wei et al. 2001). This hypothesis is also in line with that of Poirazi & Mel (2001), who have proposed that synaptic plasticity during development favours spatial clustering of synaptic inputs that carry similar or paired information. Spatial clustering of synaptic afferents within a given cortical layer is well established in sensory cortical areas (Szentagothai, 1978; Gilbert, 1983; Zeki & Shipp, 1988), and also in prefrontal cortex (Goldman & Nauta, 1976; Schwartz & Goldman-Rakic, 1991). Dendrite-specific clustering of glutamatergic synaptic afferents, together with the integration power of a single dendritic branch, could qualify as distinguishing features of the neocortical pyramidal cell.

Acknowledgments

We thank James E. Swain for comments on the manuscript. S.A. is grateful to Guy Major for valuable discussions and suggestions. This work was supported by NIH grants MH063503 and MH044866.

Supplementary material

The online version of this paper can be accessed online at:

10.1113/jphysiol.2004.061416

http://jp.physoc.org/cgi/content/full/jphysiol.2004.061416V1/DC1 and contains supplementary material consisting of two figures and a movie.

This material can also be found at: http://blackwellpublishing.com/products/journals/tjp/tjp345/tjp345sm.htm

References

  1. Antic SD. Action potentials in basal and oblique dendrites of rat neocortical pyramidal neurons. J Physiol. 2003;550:35–50. doi: 10.1113/jphysiol.2002.033746. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Antic S, Major G, Zecevic D. Fast optical recordings of membrane potential changes from dendrites of pyramidal neurons. J Neurophysiol. 1999;82:1615–1621. doi: 10.1152/jn.1999.82.3.1615. [DOI] [PubMed] [Google Scholar]
  3. Archie KA, Mel BW. A model for intradendritic computation of binocular disparity. Nat Neurosci. 2000;3:54–63. doi: 10.1038/71125. [DOI] [PubMed] [Google Scholar]
  4. Branchereau P, Van Bockstaele EJ, Chan J, Pickel VM. Pyramidal neurons in rat prefrontal cortex show a complex synaptic response to single electrical stimulation of the locus coeruleus region: evidence for antidromic activation and GABAergic inhibition using in vivo intracellular recording and electron microscopy. Synapse. 1996;22:313–331. doi: 10.1002/(SICI)1098-2396(199604)22:4<313::AID-SYN3>3.0.CO;2-E. [DOI] [PubMed] [Google Scholar]
  5. Cauller LJ, Kulics AT. The neural basis of the behaviorally relevant N1 component of the somatosensory-evoked potential in SI cortex of awake monkeys: evidence that backward cortical projections signal conscious touch sensation. Exp Brain Res. 1991;84:607–619. doi: 10.1007/BF00230973. [DOI] [PubMed] [Google Scholar]
  6. Compte A, Brunel N, Goldman-Rakic PS, Wang XJ. Synaptic mechanisms and network dynamics underlying spatial working memory in a cortical network model. Cereb Cortex. 2000;10:910–923. doi: 10.1093/cercor/10.9.910. [DOI] [PubMed] [Google Scholar]
  7. Connors BW, Gutnick MJ. Intrinsic firing patterns of diverse neocortical neurons. Trends Neurosci. 1990;13:99–104. doi: 10.1016/0166-2236(90)90185-d. [DOI] [PubMed] [Google Scholar]
  8. Cowan RL, Sesack SR, Van Bockstaele EJ, Branchereau P, Chain J, Pickel VM. Analysis of synaptic inputs and targets of physiologically characterized neurons in rat frontal cortex: combined in vivo intracellular recording and immunolabeling. Synapse. 1994;17:101–114. doi: 10.1002/syn.890170206. [DOI] [PubMed] [Google Scholar]
  9. Cowan RL, Wilson CJ. Spontaneous firing patterns and axonal projections of single corticostriatal neurons in the rat medial agranular cortex. J Neurophysiol. 1994;71:17–32. doi: 10.1152/jn.1994.71.1.17. [DOI] [PubMed] [Google Scholar]
  10. Deuchars J, West DC, Thomson AM. Relationships between morphology and physiology of pyramid-pyramid single axon connections in rat neocortex in vitro. J Physiol. 1994;3:423–435. doi: 10.1113/jphysiol.1994.sp020262. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Euler T, Detwiler PB, Denk W. Directionally selective calcium signals in dendrites of starburst amacrine cells. Nature. 2002;418:845–852. doi: 10.1038/nature00931. [DOI] [PubMed] [Google Scholar]
  12. Feix JB, Kalyanaraman B. Production of singlet oxygen-derived hydroxyl radical adducts during merocyanine-540-mediated photosensitization: analysis by ESR-spin trapping and HPLC with electrochemical detection. Arch Biochem Biophys. 1991;291:43–51. doi: 10.1016/0003-9861(91)90103-p. [DOI] [PubMed] [Google Scholar]
  13. Finch EA, Augustine GJ. Local calcium signalling by inositol-1,4,5-trisphosphate in Purkinje cell dendrites. Nature. 1998;396:753–756. doi: 10.1038/25541. [DOI] [PubMed] [Google Scholar]
  14. Funahashi S, Bruce CJ, Goldman-Rakic PS. Mnemonic coding of visual space in the monkey's dorsolateral prefrontal cortex. J Neurophysiol. 1989;61:331–349. doi: 10.1152/jn.1989.61.2.331. [DOI] [PubMed] [Google Scholar]
  15. Gilbert CD. Microcircuitry of the visual cortex. Annu Rev Neurosci. 1983;6:217–247. doi: 10.1146/annurev.ne.06.030183.001245. [DOI] [PubMed] [Google Scholar]
  16. Golding NL, Spruston N. Dendritic sodium spikes are variable triggers of axonal action potentials in hippocampal CA1 pyramidal neurons. Neuron. 1998;21:1189–1200. doi: 10.1016/s0896-6273(00)80635-2. [DOI] [PubMed] [Google Scholar]
  17. Goldman MS, Levine JH, Major G, Tank DW, Seung HS. Robust Persistent Neural Activity in a Model Integrator with Multiple Hysteretic Dendrites per Neuron. Cereb Cortex. 2003;13:1185–1195. doi: 10.1093/cercor/bhg095. [DOI] [PubMed] [Google Scholar]
  18. Goldman PS, Nauta WJ. Autoradiographic demonstration of a projection from prefrontal association cortex to the superior colliculus in the rhesus monkey. Brain Res. 1976;116:145–149. doi: 10.1016/0006-8993(76)90256-0. [DOI] [PubMed] [Google Scholar]
  19. Goldman-Rakic PS. Cellular basis of working memory. Neuron. 1995;14:477–485. doi: 10.1016/0896-6273(95)90304-6. [DOI] [PubMed] [Google Scholar]
  20. Helmchen F, Svoboda K, Denk W, Tank DW. In vivo dendritic calcium dynamics in deep-layer cortical pyramidal neurons. Nat Neurosci. 1999;2:989–996. doi: 10.1038/14788. [DOI] [PubMed] [Google Scholar]
  21. Kerr JND, Plenz D. Action potential timing determines dendritic calcium during striatal up-states. J Neurosci. 2004;24:877–885. doi: 10.1523/JNEUROSCI.4475-03.2004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Koch C, Wang HT, Mathur B. Computing motion in the primate's visual system. J Exp Biol. 1989;146:115–139. doi: 10.1242/jeb.146.1.115. [DOI] [PubMed] [Google Scholar]
  23. Koester HJ, Sakmann B. Calcium dynamics in single spines during coincident pre- and postsynaptic activity depend on relative timing of back-propagating action potentials and subthreshold excitatory postsynaptic potentials. Proc Natl Acad Sci U S A. 1998;95:9596–9601. doi: 10.1073/pnas.95.16.9596. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Krieg M. Singlet oxygen production and fluorescence yields of merocyanine 540: a comparative study in solution and model membrane systems. Biochim Biophys Acta. 1992;2:333–335. doi: 10.1016/0005-2736(92)90212-5. [DOI] [PubMed] [Google Scholar]
  25. Kritzer MF, Goldman-Rakic PS. Intrinsic circuit organization of the major layers and sublayers of the dorsolateral prefrontal cortex in the rhesus monkey. J Comp Neurol. 1995;359:131–143. doi: 10.1002/cne.903590109. [DOI] [PubMed] [Google Scholar]
  26. Lampl I, Reichova I, Ferster D. Synchronous membrane potential fluctuations in neurons of the cat visual cortex. Neuron. 1999;22:361–374. doi: 10.1016/s0896-6273(00)81096-x. [DOI] [PubMed] [Google Scholar]
  27. Larkman AU. Dendritic morphology of pyramidal neurones of the visual cortex of the rat. III. Spine distributions. J Comp Neurol. 1991;306:332–343. doi: 10.1002/cne.903060209. [DOI] [PubMed] [Google Scholar]
  28. Lasser-Ross N, Miyakawa H, Lev-Ram V, Young SR, Ross WN. High time resolution fluorescence imaging with a CCD camera. J Neurosci Meth. 1991;36:253–261. doi: 10.1016/0165-0270(91)90051-z. [DOI] [PubMed] [Google Scholar]
  29. Lewis BL, O'Donnell P. Ventral tegmental area afferents to the prefrontal cortex maintain membrane potential ‘up’ states in pyramidal neurons via D(1) dopamine receptors. Cereb Cortex. 2000;10:1168–1175. doi: 10.1093/cercor/10.12.1168. [DOI] [PubMed] [Google Scholar]
  30. Llano I, Dreessen J, Kano M, Konnerth A. Intradendritic release of calcium induced by glutamate in cerebellar Purkinje cells. Neuron. 1991;7:577–583. doi: 10.1016/0896-6273(91)90370-f. [DOI] [PubMed] [Google Scholar]
  31. London M, Schreibman A, Hausser M, Larkum ME, Segev I. The information efficacy of a synapse. Nat Neurosci. 2002;5:332–340. doi: 10.1038/nn826. [DOI] [PubMed] [Google Scholar]
  32. Luscher C, Streit J, Quadroni R, Luscher HR. Action potential propagation through embryonic dorsal root ganglion cells in culture. I. Influence of the cell morphology on propagation properties. J Neurophysiol. 1994;72:622–633. doi: 10.1152/jn.1994.72.2.622. [DOI] [PubMed] [Google Scholar]
  33. MacDermott AB, Mayer ML, Westbrook GL, Smith SJ, Barker JL. NMDA-receptor activation increases cytoplasmic calcium concentration in cultured spinal cord neurones. Nature. 1986;321:519–522. doi: 10.1038/321519a0. [DOI] [PubMed] [Google Scholar]
  34. Markram H, Helm PJ, Sakmann B. Dendritic calcium transients evoked by single back-propagating action potentials in rat neocortical pyramidal neurons. J Physiol. 1995;485:1–20. doi: 10.1113/jphysiol.1995.sp020708. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Markram H, Lubke J, Frotscher M, Roth A, Sakmann B. Physiology and anatomy of synaptic connections between thick tufted pyramidal neurones in the developing rat neocortex. J Physiol. 1997;500:409–440. doi: 10.1113/jphysiol.1997.sp022031. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Markram H, Sakmann B. Calcium transients in dendrites of neocortical neurons evoked by single subthreshold excitatory postsynaptic potentials via low-voltage-activated calcium channels. Proc Natl Acad Sci U S A. 1994;91:5207–5211. doi: 10.1073/pnas.91.11.5207. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Miller EK, Erickson CA, Desimone R. Neural mechanisms of visual working memory in prefrontal cortex of the macaque. J Neurosci. 1996;16:5154–5167. doi: 10.1523/JNEUROSCI.16-16-05154.1996. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Nakamura T, Barbara JG, Nakamura K, Ross WN. Synergistic release of Ca2+ from IP3-sensitive stores evoked by synaptic activation of mGluRs paired with backpropagating action potentials. Neuron. 1999;24:727–737. doi: 10.1016/s0896-6273(00)81125-3. [DOI] [PubMed] [Google Scholar]
  39. Oakley JC, Schwindt PC, Crill WE. Initiation and propagation of regenerative Ca(2+)-dependent potentials in dendrites of layer 5 pyramidal neurons. J Neurophysiol. 2001a;86:503–513. doi: 10.1152/jn.2001.86.1.503. [DOI] [PubMed] [Google Scholar]
  40. Oakley JC, Schwindt PC, Crill WE. Dendritic calcium spikes in layer 5 pyramidal neurons amplify and limit transmission of ligand-gated dendritic current to soma. J Neurophysiol. 2001b;86:514–527. doi: 10.1152/jn.2001.86.1.514. [DOI] [PubMed] [Google Scholar]
  41. Poirazi P, Mel BW. Impact of active dendrites and structural plasticity on the memory capacity of neural tissue. Neuron. 2001;29:779–796. doi: 10.1016/s0896-6273(01)00252-5. [DOI] [PubMed] [Google Scholar]
  42. Rapp M, Yarom Y, Segev I. Modeling back propagating action potential in weakly excitable dendrites of neocortical pyramidal cells. Proc Natl Acad Sci U S A. 1996;93:11985–11990. doi: 10.1073/pnas.93.21.11985. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Ross WN, Arechiga H, Nicholls JG. Optical recording of calcium and voltage transients following impulses in cell bodies and processes of identified leech neurons in culture. J Neurosci. 1987;7:3877–3887. doi: 10.1523/JNEUROSCI.07-12-03877.1987. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Sanchez-Vives MV, McCormick DA. Cellular and network mechanisms of rhythmic recurrent activity in neocortex. Nat Neurosci. 2000;3:1027–1034. doi: 10.1038/79848. [DOI] [PubMed] [Google Scholar]
  45. Schiller J, Major G, Koester HJ, Schiller Y. NMDA spikes in basal dendrites of cortical pyramidal neurons. Nature. 2000;404:285–289. doi: 10.1038/35005094. [DOI] [PubMed] [Google Scholar]
  46. Schiller J, Schiller Y, Clapham DE. NMDA receptors amplify calcium influx into dendritic spines during associative pre- and postsynaptic activation. Nat Neurosci. 1998;1:114–118. doi: 10.1038/363. [DOI] [PubMed] [Google Scholar]
  47. Schiller J, Schiller Y, Stuart G, Sakmann B. Calcium action potentials restricted to distal apical dendrites of rat neocortical pyramidal neurons. J Physiol. 1997;505:605–616. doi: 10.1111/j.1469-7793.1997.605ba.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Schwartz ML, Goldman-Rakic PS. Prenatal specification of callosal connections in rhesus monkey. J Comp Neurol. 1991;307:144–162. doi: 10.1002/cne.903070113. [DOI] [PubMed] [Google Scholar]
  49. Shepherd GM. The neuron doctrine: a revision of functional concepts. Yale J Biol Med. 1972;45:584–599. [PMC free article] [PubMed] [Google Scholar]
  50. Shu YS, Hasenstaub A, McCormick DA. Turning on and off recurrent balanced cortical activity. Nature. 2003;423:288–293. doi: 10.1038/nature01616. [DOI] [PubMed] [Google Scholar]
  51. Steriade M, Nunez A, Amzica F. A novel slow (< 1 Hz) oscillation of neocortical neurons in vivo: depolarizing and hyperpolarizing components. J Neurosci. 1993a;13:3252–3265. doi: 10.1523/JNEUROSCI.13-08-03252.1993. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Steriade M, Nunez A, Amzica F. c (< 1 Hz) neocortical oscillation and other sleep rhythms of the electroencephalogram. J Neurosci. 1993b;13:3266–3283. doi: 10.1523/JNEUROSCI.13-08-03266.1993. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Stuart GJ, Sakmann B. Active propagation of somatic action potentials into neocortical pyramidal cell dendrites. Nature. 1994;367:69–72. doi: 10.1038/367069a0. [DOI] [PubMed] [Google Scholar]
  54. Szentagothai J. The neuron network of the cerebral cortex: a functional interpretation. P Roy Soc Lond B Bio. 1978;201:219–248. doi: 10.1098/rspb.1978.0043. [DOI] [PubMed] [Google Scholar]
  55. Takechi H, Eilers J, Konnerth A. A new class of synaptic response involving calcium release in dendritic spines. Nature. 1998;396:757–760. doi: 10.1038/25547. [DOI] [PubMed] [Google Scholar]
  56. Wei DS, Mei YA, Bagal A, Kao JP, Thompson SM, Tang CM. Compartmentalized and binary behavior of terminal dendrites in hippocampal pyramidal neurons. Science. 2001;293:2272–2275. doi: 10.1126/science.1061198. [DOI] [PubMed] [Google Scholar]
  57. Williams GV, Goldman-Rakic PS. Modulation of memory fields by dopamine D1 receptors in prefrontal cortex. Nature. 1995;376:572–575. doi: 10.1038/376572a0. [DOI] [PubMed] [Google Scholar]
  58. Wilson CJ, Groves PM. Spontaneous firing patterns of identified spiny neurons in the rat neostriatum. Brain Res. 1981;220:67–80. doi: 10.1016/0006-8993(81)90211-0. [DOI] [PubMed] [Google Scholar]
  59. Wilson CJ, Kawaguchi Y. The origins of two-state spontaneous membrane potential fluctuations of neostriatal spiny neurons. J Neurosci. 1996;16:2397–2410. doi: 10.1523/JNEUROSCI.16-07-02397.1996. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Zecevic D. Multiple spike-initiation zones in single neurons revealed by voltage-sensitive dyes. Nature. 1996;381:322–325. doi: 10.1038/381322a0. [DOI] [PubMed] [Google Scholar]
  61. Zeki S, Shipp S. The functional logic of cortical connections. Nature. 1988;335:311–317. doi: 10.1038/335311a0. [DOI] [PubMed] [Google Scholar]

Articles from The Journal of Physiology are provided here courtesy of The Physiological Society

RESOURCES