Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2012 Jun 21.
Published in final edited form as: Methods Mol Biol. 2006;337:127–137. doi: 10.1385/1-59745-095-2:127

Juxtacellular Labeling and Chemical Phenotyping of Extracellularly Recorded Neurons In Vivo

Glenn M Toney, Lynette C Daws
PMCID: PMC3380431  NIHMSID: NIHMS295789  PMID: 16929944

Summary

Extracellular recording of the action potential discharge of individual neurons has been an indispensable electrophysiological method for more than 50 yr. Although it requires relatively modest instrumentation, extracellular recording nevertheless provides critically important information concerning the patterning of intercellular communication in the nervous system. In 1996, Didier Pinault described “juxtacellular labeling” as “a novel and very effective single-cell labeling method” for revealing the morphology of extracellularly recorded neurons. Of particular interest for neuroscience is that juxtacellular labeling can be combined with immunocytochemistry and in situ hybridization histochemistry to reveal new and exciting information concerning the chemical phenotype of neurons whose electrophysiological properties have been characterized in vivo. By providing investigators with a means to “match” functional information from electrophysiological recordings with morphological and protein/gene expression data at the level of the single neuron, juxtacellular labeling has opened a new era in neuroscience research, one that holds the promise of an accelerated pace of discovery.

Keywords: Confocal microscopy, cellular morphology, electrophysiology, extracellular recording, histochemistry, immunocytochemistry, in situ hybridization

1. Introduction

In the mammalian nervous system, neurons communicate (often over long distances) by generating patterns of action potentials. Ion channels lie at the heart of this process. Indeed, individual neurons in the intact nervous system express a host of ion channels with permeability and gating characteristics that determine the shape and patterning of action potentials. Recording the discharge of individual neurons has been at the forefront of neuroscience research for many years, and this core methodology has been described at length elsewhere (1-3).

In its simplest form, extracellular single-unit recording involves introducing a glass, metal, or carbon fiber microelectrode with relatively low tip resistance (usually 5–40 MΩ) into the extracellular space adjacent to a neuron of interest. Action potentials of individual neurons are recorded as transient changes in electrode potential with the aid of a low-noise alternating current (AC)-coupled amplifier. Signals are referenced to an extracellularly located electrode comprised of a nonpolarizing metal composite such as Ag/AgCl. In extracellular recording mode, voltages at the microelectrode are typically small (10–500 μV) owing to the low resistivity of extracellular fluid. Thus, considerable amplification is generally employed so that recorded signals effectively utilize the entire input voltage range of most commercially available analog-to-digital converters (ADCs; ±5 V). This is used in combination with a computer and data acquisition software to facilitate data storage, retrieval, and analysis.

The enduring utility of in vivo extracellular single-neuron recording can be attributed to the importance of determining the discharge response of neurons to specific afferent inputs and to the many accessory methods that have been developed over the years. For example, antidromic mapping and collision testing can be used to identify the location and branching pattern of the axon terminals of individual cells (4). Multibarrel glass electrodes can also be used to couple extracellular recording with pressure ejection or microiontophoresis of compounds such as receptor selective ligands (5). Such studies allow determination of the contribution of specific neurotransmitters and receptors in controlling tonic activity and synaptically evoked neuronal responses. The behavior and organization of neural networks can be studied by simultaneously recording the discharge of multiple cells using an array of electrodes and sophisticated statistical correlation algorithms (6). Of particular interest is that the latter studies can be achieved in conscious, freely behaving animals (7).

Here, we focus on a relatively new methodology termed juxtacellular labeling, which allows histological staining and visualization of individual neurons whose discharge behavior has been electrophysiologically characterized. This methodology was developed by Didier Pinault and was introduced in 1996 (8). Anodal (positive) current pulses are passed through a glass recording electrode, which is filled with a solution containing biotinamide. The current pulses electrostatically repel positively charged biotinamide molecules, expelling them from the electrode. Because of the proximity of the electrode tip to the membrane of the recorded neuron, current pulses effectively cause single-cell electroporationa transient opening created in the plasma membrane through which biotinamide molecules from the electrode gain entry into the cell interior, thus filling the cytoplasm of the soma and often the proximal dendrites (for review, see ref. 9).

When combined with antidromic activation, juxtacellular labeling allows the precise location and morphology of an individual neuron to be determined along with (in limited situations) its downstream connectivity. When further combined with immunocytochemistry or in situ hybridization histochemistry, the experimenter is able to determine the connectivity, morphology, and chemical phenotype of an individually recorded and electrophysiologically characterized neuron. Indeed, in vivo studies have used this experimental approach to record the discharge of individual neurons in brain regions including the thalamus (8), medial septum (10), dorsal raphe nucleus (DRN) (11), rostral ventrolateral medulla (9,12,13), cerebellum (14), cochlear nucleus (15), and retrotrapezoid nucleus (16) and, in the context of a single experiment, to determine the somatodendritic morphology of recorded cells along with details concerning their neurotransmitter content (9,11,12), expression of receptor/transporter protein/messenger (m)RNA (9,10), cytosolic enzymes (9,12), and ion channel subunit mRNA (13).

2. Materials

  1. Borosilicate glass (1.5 mm od, 0.86 mm id, World Precision Instruments, Inc., Sarasota, FL).

  2. Narishige vertical puller (model PE-2, Narishige International USA, Inc., East Meadow, NY).

  3. Axoclamp 2B amplifier with HS-2A x0.1LU headstage (Axon Instruments, Inc., Foster City, CA).

  4. AC preamplifier (Grass, model P15D, Astromed, Inc., Grass Instruments Division, West Warwick, RI).

  5. ADC (model 1401plus/micro1401, Cambridge Electronic Design, Inc., Cambridge, UK).

  6. Window discriminator (model 121, World Precision Instruments).

  7. Data acquisition/analysis software (Spike2, v4.31, Cambridge Electronic Design, Inc.).

  8. Pulse generator (model PulseMaster, World Precision Instruments).

  9. Sodium acetate, sodium chloride (Sigma, St. Louis, MO).

  10. Biotinamide (Molecular Probes, Eugene, OR) or neurobiotin (N-[2 aminoethyl] biotinamide HCl, Vector Laboratories, Burlingame, CA).

  11. Anesthesia for rats: cocktail containing urethane (1.0 g/mL) and α-chloralose (0.1 g/mL) at a dose of 800 mg/kg urethane and 80 mg/kg chloralose ip. (Combine 1 g urethane with 0.1 mL normal saline and warm to approx 40°C, then add chloralose and vortex/stir until dissolved.)

  12. 0.1 M Phosphate-buffered saline (PBS).

  13. PBS containing 4% paraformaldehyde (PFA).

  14. PBS supplemented with 30% sucrose.

  15. Freezing microtome/cryostat.

  16. Polyvinylpyrrolidine (PVP-40) cryoprotectant: 30 g sucrose, 30 mL ethylene glycol, and 1.0 g PVP-40 brought to 100 mL with PBS.

  17. 100 mM Tris-buffered saline (TBS)

  18. Triton X-100.

  19. Streptavidin-Texas red or streptavidin-FITC (fluorescein isothiocyanate) (Jackson ImmunoResearch Laboratories, West Grove, PA).

  20. Avidin-peroxidase conjugate (ABC Vectastain Kit, Vector).

  21. 3,3’-Diaminobenzidine tetrahydrochloride (DAB), hydrogen peroxide, and nickel sulfate hexahydrate (Sigma).

  22. Normal goat serum (NGS; unless primary antibody is from goat, in which case use donkey serum).

  23. Primary antibody of interest (e.g., monoclonal antitryptophan hydroxylase antibody). For immunohistochemistry, this antibody may be conjugated to a fluorophore, such as FITC or Texas red. Alternatively, an appropriate secondary antibody conjugated to FITC (Caltag Laboratories, Burlingame, CA, cat. no. M32601) or Texas red may be used to identify the primary antibody.

  24. Superfrost Plus slides (Fisher, Hampton, NH).

  25. Ethanol (70, 95, and 100%)

  26. Xylene.

  27. Cytoseal 60 (Fisher).

3. Methods

3.1. In Vivo Extracellular Single-Unit Recording

  1. Borosilicate glass microelectrodes are pulled to a fine-tip diameter (1–2 μm) using a Narishige vertical puller. The tips are blunted by making contact with the broken flat surface of a glass rod positioned under a microscope. The final tip resistance is typically 15–40 MΩ.

  2. Microelectrodes are filled with 0.5 M sodium acetate supplemented with 5% biotinamide.

  3. We obtain recordings with a direct current (DC) intracellular amplifier. We prefer the Axoclamp 2B in bridge mode (Axon). We then pass the ×10 output of the DC amplifier to a battery-powered model P15D AC preamplifier (Astro-Med). The amplifier has half-amplitude frequency filters that we usually set to a bandpass of 0.3–3.0 kHz. Use of a 60-Hz notch filter can also be advantageous.

  4. The output of the AC amplifier is then sent to an ADC (model 1401plus, Cambridge Electronic Design) and to a window discriminator (model 121, World Precision Instruments). The window discriminator multiplex output is sent to an analog oscilloscope and to the ADC to monitor cell discharge. We use Spike2 (v4.31) data acquisition and analysis software (Cambridge Electronic Design). Data is stored on a computer.

3.2. Juxtacellular Labeling

  1. The juxtacellular labeling procedure has been described elsewhere (8,9). Once a single-unit recording is obtained, as evident by spikes of uniform shape and amplitude, and physiological studies have been completed, juxtacellular labeling is initiated (see Fig. 1A).

  2. The Axoclamp 2B amplifier has convenient front-mounted step command switches that facilitate juxtacellular labeling by allowing accurate and rapid adjustment of the amplitude of current pulses delivered through the recording electrode (see step 3). When using the standard HS-2A x0.1LU head stage, the full-scale output is ±19.99 nA, which is typically more than adequate for juxtacellular labeling (see Note 1).

  3. Anodal (positive) current pulses are delivered through the recording electrode. The standard approach is to deliver 200-ms pulses with a 50% duty cycle (see Fig. 1B). It is most convenient to trigger current pulses with an external pulse generator such as the PulseMaster (World Precision Instruments). Alternatively, the digital-to-analog output available with most digital data acquisitions systems can be programmed for the same purpose.

  4. The amplitude of current pulses is gradually increased while continuously monitoring the discharge of the individually recorded neuron. The goal is to entrain the discharge of the recorded neuron to the timing of current pulses (see Fig. 1C). The amplitude of current pulses required to entrain discharge varies. Factors likely to influence this include specific properties of the recorded neurons as well as properties of the recording electrode, such as tip resistance, diameter, and overall shape.

  5. In our experience, entrainment of discharge among neurons of the DRN is successful in the majority of cases at current amplitudes ranging from 2 to 8 nA. Occasionally, entrainment may require that the recording electrode be moved in small increments toward the recorded neuron. Depending on properties of the recorded cell, an increase in signal noise may occur as an indication that the microelectrode has reached the “juxtacellular” location. In this position, current pulses of the appropriate amplitude will most often result in discharge entrainment. During entrainment, electroporation of the cell membrane will result in filling of the cytoplasm with biotinamide. It is sometimes observed that cells that resist entrainment will succumb if the amplitude of current pulses is increased to 15–20 nA for a few cycles and then quickly reduced (see Note 2).

  6. After discharge becomes entrained to the timing of current pulses, successful filling will typically be achieved within a short time, usually 30 s to 5–10 min. How long entrainment must be maintained for adequate filling to occur depends on several factors, including the size and morphology of the cell, electrode shape/resistance, and concentration of biotinamide in the recording electrode. Another factor that may influence the efficacy of entrainment and cell filling is the nature of the microelectrode filling solution. Some investigators report improved results when electrodes are filled with 0.5 M sodium acetate compared to sodium chloride. Presumably, this reflects the valence or mobility of the counteranion in solution with biotinamide.

  7. Once entrainment has been maintained for a sufficiently long period, the amplitude of current pulses is reduced, and pulse delivery is terminated. It is advisable to record spontaneous cell discharge for several minutes thereafter to assess the “health” of the recorded neuron following the entrainment/labeling procedure. At this point, the recording electrode can be removed. Depending on the goal of the experiment, it may be advisable to delay perfusion fixation of the animal for up to several hours to maximize diffusion/distribution of biotinamide throughout the cell interior. The last is particularly important if the morphology of dendrites is to be examined and if local connectivity is to be observed.

Fig. 1.

Fig. 1

Open in a new tab

(A) Single-unit recording of the spontaneous discharge of a slow-firing (0.5-Hz) neuron in the DRN of a C57Bl/6 mouse. (B) Anodal (positive) current pulses (200-ms 50% duty cycle) were passed through the recording electrode, and the amplitude was adjusted to 8 nA. (C) The discharge of the recorded neuron became entrained to the timing of current pulses. Note that current-switching artifacts are clipped to emphasis the entrained discharge. Voltage calibration is postamplification.

3.3. Recovering Biotinamide-Filled Neurons

  1. After recording and juxtacellularly labeling a neuron (see Subheading 3.2.), the still-anesthetized rat is perfused transcardially with 100 mL PBS followed by perfusion with 330 mL chilled (4°C) PBS containing 4% PFA to fix the brain tissue. The rate of perfusion is approx 33 mL/min. For fixation of mouse brain, the volume and rate of transcardiac perfusion are reduced to 10% of the values for the rat.

  2. The brain is removed and postfixed in PFA for 4–24 h. To ensure thorough cryoprotection, brains are then transferred to 30% sucrose-PBS until the tissue block sinks (~48 h). At this time, tissue can either be placed in frozen storage or cut into sections 30-μm thick. If stored, brains should be frozen in chilled isopentane before placing them in a freezer. Isopentane is placed in a glass beaker surrounded by crushed dry ice until chilled (~15 min), and brains are then immersed until fully frozen. We wrap brains in aluminum foil and place them directly in a freezer at −80°C. If brains are immediately sectioned, a freezing microtome/cryostat is used at −20°C. After cutting, sections can be stored at −20°C in PVP-40 cryoprotectant prior to processing for recovery of juxtacellular labeling and immunostaining.

  3. Tissue sections stored in cryoprotectant at −20°C are warmed to room temperature (30 min) and rinsed (five times for 5 min/rinse) in TBS containing 0.1% Triton X-100.

  4. Biotinamide-filled cells are recovered by reacting tissue with streptavidin in TBS containing 0.1% Triton X-100. Streptavidin can be conjugated to a wide variety of fluorochromes. We have used streptavidin-Texas red or streptavidin-FITC at a concentration of 1:200. Alternatively, soma and dendritic morphology can be analyzed by recovering biotinamide-filled cells using a standard avidin-peroxidase/diaminobenzadine reaction with metal (Ni2+) intensification. The latter has the decided advantage that the resultant staining is blue/black and is effectively permanent.

  5. Sections are incubated in TBS containing streptavidin-Texas red (1:200) for 1 h at room temperature and then for 24 h at 4°C with mild agitation on a rocker table. We carry out the reaction in a 1.5-mL microcentrifuge tube to minimize the amount of fluorochrome used. Sections are covered with foil to minimize photobleaching.

  6. Sections are warmed to room temperature and rinsed in TBS (five times for 5 min/rinse), mounted on Superfrost Plus slides (Fisher), and air-dried overnight while covered with aluminum foil.

  7. Sections are dehydrated in a series of ethanol (3 min each in 70, 95, 95, 100%), cleared (defatted) in xylene (2–3 min each), and cover slipped with Cytoseal 60 (Fisher) (see Fig. 2).

Fig. 2.

Fig. 2

Open in a new tab

A wide-field fluorescent (Texas red) micrograph reveals the juxtacellularly labeled neuron in the DRN for which the entrainment may be seen in Fig. 1. Note that the soma and proximal dendrites are filled with biotinamide. CA, cerebral aqueduct.

3.4. Colocalization of Cellular Proteins by Immunocytochemistry

  1. Immunocytochemical staining to colocalize immunoreactive proteins within juxtacellularly labeled neurons can be performed before or after recovery of biotinamide-filled cells. The optimal sequence of staining usually depends on whether juxtacellularly labeled cells are recovered using a fluorescent marker or a peroxidase method as previously described. If different fluorochromes are used for recovering the juxtacellularly labeled cell and for immunostaining, the sequence used for the two staining steps will usually depend on their relative intensities. The intensity of fluorescence will depend on a number of factors, including the degree of biotinamide filling (juxtacellular label), the density of protein expression, antibody affinities, and the type of amplification system used for staining.

  2. As noted in the Subheading 1., studies to date have successfully combined juxtacellular labeling with immunocytochemical staining to localize a number of cellular proteins within electrophysiologically characterized central neurons. Here, we present our methodology for immunostaining for tryptophan hydroxylase (TPH), the rate-limiting enzyme in the synthesis of serotonin (5HT). We have used this approach to neurochemically phenotype serotonergic neurons in the DRN of C57BL/6 mice (see Fig. 3).

  3. First, tissue sections through the DRN (stored in PPV-40 cryoprotectant) are warmed to room temperature (30 min) and placed in chambers with a mesh bottom to facilitate transfer to various solutions in the staining protocol.

  4. Sections are rinsed in TBS (five times for 5 min/rinse) and incubated in a monoclonal primary anti-TPH antibody (Sigma, cat. no. T0678) diluted 1:500 in TBS containing 1% NGS and 0.1% Triton X-100 (see Note 3).

  5. Sections are incubated with gentle agitation at room temperature for 1 h and at 4°C for 48 h.

  6. Next, tissue sections are again warmed to room temperature and rinsed in TBS (five times for 5 min/rinse).

  7. Sections are then incubated with gentle agitation in goat antimouse immunoglobulin G3 secondary antibody conjugated to FITC (Caltag Laboratories, cat. no. M32601) for 1 h at room temperature and 6–9 h at 4°C. Sections are covered with aluminum foil to minimize photobleaching.

  8. If sections were processed earlier for recovery of juxtacellularly labeled cells, the procedure is now complete, and sections can be rinsed in TBS (five times for 5 min/rinse), cleared, and cover slipped as indicated in Subheading 3.3.4.

  9. If sections have not yet been processed for recovery of the juxtacellular label, the procedure may now be performed by immersing sections in streptavidin-Texas red diluted 1:200 in TBS containing 0.1% Triton X-100 and 1% NGS and incubating sections for 1 h at room temperature and for 24 h at 4°C.

  10. Sections are then rinsed in TBS, cleared, mounted on slides, and cover slipped as previously described.

  11. Sections can then be viewed under an appropriately equipped fluorescent microscope (wide field or confocal) to determine the presence of double-labeled neurons (see Fig. 3).

Fig. 3.

Fig. 3

Open in a new tab

Top, laser scanning confocal photomicrograph (×10) showing colocalization of tryptophan hydroxylase (TPH) immunofluorescence (green) in a juxtacellularly labeled neuron (red) in the DRN of a C57Bl/6 mouse. Arrow indicates the recorded neuron. Bottom, high-magnification (×60) image of juxtacellularly labeled DRN neuron (left), TPH immunofluorescence (center), and a merged image showing colocalization at a 0.6-μm z-plane resolution. CA, cerebral aqueduct.

Acknowledgments

We thank Mr. Alfredo Calderon for excellent technical assistance. This work was supported by National Institute of Health grants HL71645, HL76312 (G. M. T.), and MH64489 (L. C. D.).

Footnotes

1

We have successfully implemented juxtacellular labeling with a variety of instruments. It is possible to use a relatively inexpensive DC amplifier like the model 767 Electrometer from World Precision Instruments; it has a unitary gain headstage. Because this unit also has only 10x and 50x output, delivery of even small current pulses (<10 nA) will saturate our AC amplifier. To avoid this, we impose an attenuating (x0.1) amplifier/isolation unit (model SIU5A, Grass Instruments) before the signal is passed to the AC amplifier. It is a simple matter then to use a pulse generator to drive the external input to the current isolator circuit of the amplifier to trigger and grade the amplitude of current pulses through the recording microelectrode. The amplitude of current pulses can be readily monitored by simply calibrating a channel of the computer data acquisition software or an oscilloscope.

2

Labeling multiple neurons: it is possible for the juxtacellular labeling method to result in filling of more than one cell. The incidence of this is generally low, particularly when care is taken to monitor cell discharge continuously to ensure that only the individually recorded neuron becomes entrained by current pulses. The frequency of labeling multiple neurons in the DRN of mice in our experience is low (<5%). Guyenet et al. (9) also reported that multiple-cell filling is infrequent in the rostral ventrolateral medulla. As these investigators noted, however, the frequency of labeling multiple cells may increase in brain regions where the packing density of neurons is high.

3

High background staining: the TPH immunocytochemical staining procedure described for mice may be improved because of more recent availability of antimouse primary antibodies raised in nonmurine species. Alternatively, immunostaining for serotonin may be performed with excellent results.

References

  • 1.Aston-Jones GS, Siggins GR. Electrophysiology. In: Bloom FE, Kupfer DJ, editors. Psychopharmacology: The Fourth Generation of Progress. Raven Press; New York: 1995. pp. 41–63. [Google Scholar]
  • 2.Sherman-Gold R, editor. The Axon Guide For Electrophysiology and Biophysics Labortory Techniques. Axon Instruments; Foster City, CA: 1993. Instrumentation for measuring bioelectric signals from cells; pp. 17–24. [Google Scholar]
  • 3.Sherman-Gold R, editor. The Axon Guide For Electrophysiology and Biophysics Labortory Techniques. Axon Instruments; Foster City, CA: 1993. Instrumentation for measuring bioelectric signals from cells; pp. 25–80. [Google Scholar]
  • 4.Lipski J. Antidromic activation of neurones as an analytic tool in the study of the central nervous system. J Neurosci Methods. 1981;4:1–32. doi: 10.1016/0165-0270(81)90015-7. [DOI] [PubMed] [Google Scholar]
  • 5.Stone TW. Microiontophoresis and Pressure Ejection. Wiley; New York: 1985. [Google Scholar]
  • 6.Blanche TJ, Spacek MA, Hetke JF, Swindale NV. Polytrodes: high density silicon electrode arrays for large scale multiunit recording. J Neurophysiol. 2004 doi: 10.1152/jn.01023.2004. Epub. [DOI] [PubMed] [Google Scholar]
  • 7.Super H, Roelfsema PR. Chronic multiunit recordings in behaving animals: advantages and limitations. Prog Brain Res. 2005;147:263–282. doi: 10.1016/S0079-6123(04)47020-4. [DOI] [PubMed] [Google Scholar]
  • 8.Pinault D. A novel single-cell staining procedure performed in vivo under electrophysiological control: morpho-functional features of juxtacellularly labeled thalamic cells and other central neurons with biocytin or neurobiotin. J Neurosci Methods. 1996;65:113–136. doi: 10.1016/0165-0270(95)00144-1. [DOI] [PubMed] [Google Scholar]
  • 9.Guyenet PG, Stornetta RL, Weston MC, McQuiston T, Simmons JR. Detection of amino acid and peptide transmitters in physiologically identified brainstem cardiorespiratory neurons. Auton Neurosci. 2004;114:1–10. doi: 10.1016/j.autneu.2004.06.003. [DOI] [PubMed] [Google Scholar]
  • 10.Bassant MH, Simon A, Poindessous-Jazat F, Csaba Z, Epelbaum J, Dournaud P. Medial septal GABAergic neurons express the somatostatin sst2A receptor: functional consequences on unit firing and hippocampal theta. J Neurosci. 2005;25:2032–2041. doi: 10.1523/JNEUROSCI.4619-04.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Allers KA, Sharp T. Neurochemical and anatomical identification of fast- and slow-firing neurons in the rat dorsal raphe nucleus using juxtacellular labeling methods in vivo. Neuroscience. 2003;122:193–204. doi: 10.1016/s0306-4522(03)00518-9. [DOI] [PubMed] [Google Scholar]
  • 12.Schreihofer AM, Guyenet PG. Identification of C1 presympathetic neurons in rat rostral ventrolateral medulla by juxtacellular labeling in vivo. J Comp Neurol. 1997;387:524–536. doi: 10.1002/(sici)1096-9861(19971103)387:4<524::aid-cne4>3.0.co;2-4. [DOI] [PubMed] [Google Scholar]
  • 13.Washburn CP, Bayliss DA, Guyenet PG. Cardiorespiratory neurons of the rat ventrolateral medulla contain TASK-1 and TASK-3 channel mRNA. Resp Physiol Neurobiol. 2003;138:19–35. doi: 10.1016/s1569-9048(03)00185-x. [DOI] [PubMed] [Google Scholar]
  • 14.Simpson JI, Hulscher HC, Sabel-Goedknegt E, Ruigrok TJ. Between in and out: linking morphology and physiology of cerebellar cortical interneurons. Prog Brain Res. 2004;148:329–340. doi: 10.1016/S0079-6123(04)48026-1. [DOI] [PubMed] [Google Scholar]
  • 15.Arnott RH, Wallace MN, Shackleton TM, Palmer AR. Onset neurones in the anteroventral cochlear nucleus project to the dorsal cochlear nucleus. J Assoc Res Otolaryngol. 2004;5:153–170. doi: 10.1007/s10162-003-4036-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Mulkey DK, Stornetta RL, Weston MC, et al. Respiratory control by ventral surface chemoreceptor neurons in rats. Nat Neurosci. 2004;7:1360–1369. doi: 10.1038/nn1357. [DOI] [PubMed] [Google Scholar]

RESOURCES