Twenty-five male Wistar rats (300-350 g, Panum Institute, Copenhagen) were anaesthetised with halothane (Vapor, Dräger; 4 % at induction and 1.5 % during surgery) in 30 % O2-70 % N2O. Lidocaine (lignocaine; 5 mg ml−1s.c.) was used at the operation sites and at the contact spots for ear pins. Catheters were inserted in the femoral artery for recording arterial blood pressure, and in the femoral vein. Rats were tracheotomised and artificially ventilated to maintain arterial pH at 7.35-7.40, arterial partial pressure of CO2 (Pa,CO2) at approximately 39 mmHg and arterial partial pressure of O2 (Pa,O2) at approximately 100 mmHg (measured by ABL 715, Radiometer, Denmark). The head was placed in a stereotactic headholder. Cranial bones were removed over the sensory cortex (-1 to −5 mm posterior to bregma and 4-8 mm lateral to the midline) with a dental drill during continuous saline superfusion. A pool of 5 % agar in Ringer solution was made around the craniotomy for continuous superfusion with aerated (95 % air-5 % CO2) artificial cerebrospinal fluid (aCSF) at 37 °C (composition, mm: 120 NaCl, 2.8 KCl, 22 NaHCO3, 1.45 CaCl2, 1.0 Na2HPO4 and 0.876 MgCl2). The dura was carefully removed over the sensory cortex. The temperature of the animal was monitored with a rectal thermometer and kept at 37 °C by a heating pad (custom made). If animals suffered damage to the cortex during surgery they were immediately killed with an i.v. injection of air and excluded from the study.
After surgery the anaesthesia was changed to α-chloralose (1,2-O-[2,2,2-trichloro-ethylidene]-α-d-gluco-furanose), which was dissolved in saline (10 mg ml−1) and heated to 69 °C. The filtered α-chloralose was introduced i.v. as a bolus of 45 mg kg−1. Halothane and N2O was discontinued and another 15 mg kg−1 bolus was given. During the rest of the experiment 15-20 mg kg−1 was given every 20 min. The level of anaesthesia was checked continuously by observing arterial blood pressure during stimulation, and by tail pinch.
α-Chloralose is the anaesthetic of choice for brain activation studies due to minimal cardiovascular effects and preserved neuronal and vascular responses to sensory stimulation (Ngai et al. 1988; Lindauer et al. 1993; Bonvento et al. 1994). At the end of the experiment, rats were killed by an i.v. injection of air. The National Animal Ethics Committee approved all surgical and anaesthetic procedures.
Laser Doppler flowmetry
The CBF was continuously monitored by a laser Doppler flowmeter using a probe positioned 0.3-0.5 mm above the pial surface avoiding large pial vessels, as described elsewhere (Fabricius et al. 1997). The sampling depth of LDF varies as a function of the distance between the transmitting and the recording fibre (fibre separation) and of the wavelength of the laser light. We used a four-channel LDF probe (PF 415:49, Perimed AB, Jarfjalla, Sweden) holding four combinations of wavelength and fibre separation: green laser light (543 nm), fibre separation 140 μm (LDFGreen); near-infrared laser light (780 nm), fibre separations 140 μm, 250 μm and 500 μm (LDFRed(140), LDFRed(250) and LDFRed(500), respectively). The four channels of the probe were connected to two laser Doppler flowmeters (Periflux, 4001 Master, Perimed AB). The time constant was 0.2 s and the signal-processing unit used a bandwidth of 20 Hz to 12 kHz (Nilsson, 1980). The light of the recording fibre for the green laser was filtered to remove the near-infrared light and vice versa. The maximal measurement depth of the four channels in vivo has been estimated as 250 μm, 500 μm, 1000 μm and 2000 μm for LDFGreen, LDFRed(140), LDFRed(250) and LDFRed(500), respectively (Fabricius et al. 1997). The four-channel laser Doppler probe was calibrated with a motility standard and positioned corresponding to the whisker area over the somatosensory cortex avoiding the largest vessels (Chapin & Lin, 1984). After stable baseline recordings had been obtained, the probe was left for the duration of the experiment. For each channel the relative changes of CBF were calculated using the algorithm described previously (Fabricius et al. 1997). CBF changes were calculated relative to baseline, defined as the average flow for 3 s before stimulation. All CBF changes presented in text and figures represent averaged data that were transformed according to Fabricius et al. (1997). The magnitude of the CBF response was calculated as the average relative increase starting at 2 s after onset of stimulation and lasting until stimulation was stopped.
Field potential measurements
Field potentials were recorded from the cerebral cortex by single-barrel glass electrodes (pulled from capillary tubes; 1.8 mm, o.d.; 1.2 mm, i.d.; Modulohm, Denmark) filled with 2 m NaCl. The tip diameter was approximately 2 μm and the electrode impedance varied between 2 and 3 MΩ. The reference electrode consisted of a low-impedance Ag-AgCl wire resting in the neck muscles. FPs were amplified 500 times and filtered at a bandwidth of 1-2400 Hz (CyberAmp 380, Axon Instruments, Foster City, CA, USA). Neuronal signals were continuously displayed on a digital storage oscilloscope (Pintek, digital storage oscilloscope, DS-203, Taiwan). The microelectrode was lowered into the cortex using a motor-controlled micromanipulator in steps of 2 μm (custom made). The microelectrode was positioned in the somatosensory cortex corresponding to the barrel cortex at co-ordinates −3 mm posterior to bregma and −6 mm lateral to the midline at a maximal distance of 200 μm from the LDF probe (Chapin & Lin, 1984). The primary excitatory response in this study had a negative onset as it was recorded below the brain surface. The FP amplitude was measured from baseline to maximal negative peak. We evaluated that this amplitude was the most reproducible variable of neuronal activity within and between experiments. FP amplitudes were averaged for each stimulation period.
All recordings were taken after a stable baseline was obtained. On-line and off-line analysis was performed using the Spike 2 program with a 1401 plus interface (Cambridge Electronic Design, Cambridge, UK). The digital sampling rate was 10 kHz for the neuronal signals, 10 Hz for the CBF trace and 100 Hz for the blood pressure.
First, we obtained a depth profile of the FP amplitudes in response to infraorbital nerve stimulation in four animals. The purpose of this part of the study was to define the cortical depths at which the FP amplitude, and by inference neuronal activity, was maximal. The electrode was initially lowered to a depth of 2000 μm. Subsequently, the electrode was withdrawn at time intervals of 10-20 s to the following depths in micrometres: 1500, 1200, 1000, 800, 700, 600, 500, 400, 300, 200, 100, 0. At each depth the infraorbital nerve was stimulated for 50 s at 1 Hz and 2.5 mA. The largest FP amplitudes were observed at a depth of 600-800 μm. Therefore in the following experiments the electrode was positioned at this depth.
Second, we determined the optimal stimulation intensity with respect to FP and CBF amplitudes in 10 rats. In a pilot study using a stimulation frequency of 1 Hz we found that high stimulation intensities sometimes evoked cortical spreading depression that has profound and long-lasting effects on cortical function (Lauritzen, 1994). Therefore we carried out this protocol using a stimulation frequency of 3 Hz that increased CBF at lower stimulus intensities. The stimulation strength was increased from 0 to 2 mA in steps of 0.1 mA, using a frequency of 3 Hz and stimulus duration of 16 s. Data from these experiments were used to examine the relation between FP amplitudes and CBF, i.e. the activity-flow coupling.
Third, we examined the duration dependency of the CBF and FP responses by extending the period of stimulation from 8 to 16 or 32 s in random sequence using a stimulation frequency of 1 Hz and a stimulus intensity of 1.5 mA in six rats. The purpose of this part of the study was to examine whether the amplitude of the CBF response increased as a function of time as in the cerebellar cortex. Stimulus trains started every 30 s for 8 and 16 s stimulations, and every 60 s when stimulating for 32 s. CBF responses measured by LDF have been reported to vary between successive stimulations (Detre et al. 1998). Therefore, we decided to improve the signal-to-noise ratio by averaging CBF and FP data for 10 stimulation periods. Averaging of CBF and FP responses was also used in the following protocol in which we examined the influence of variations in the stimulus frequency on the responses, and in the protocol in which we examined the time course of development of CBF changes in different cortical layers.
Fourth, we examined the frequency characteristics of the FP and CBF responses in six rats. Stimulus duration was 8 s, intensity was 1.5 mA and stimulus frequencies were 0.25, 0.5, 1, 2, 3 and 5 Hz. Stimulations were repeated 10 times for each frequency with 30 s intervals between stimulus trains. Responses were averaged as described above. CBF and FP responses were similar whether moving from lower to higher frequencies or vice versa. FP amplitudes were calculated for each stimulation frequency. In addition we multiplied the amplitude with the number of stimulations to obtain the summed field potential (SFP). The relationship between FP and CBF, and SFP and CBF was examined by correlation analysis using data from the LDFRed(250).
Fifth, we carried out laminar analysis of CBF changes in real time. The purpose of this part of the study was to define the cortical depth at which CBF started and where it was maximal. The time latency was calculated from onset of stimulation to a significant CBF increase had been obtained for each LDF channel. This was defined as a CBF increase of more than two standard deviations from baseline. CBF amplitudes were calculated as described above. Data from the frequency protocol were used for this analysis.
Sixth, we examined the effect of blockade of glutamate receptors on the CBF and FP responses. The purpose of this part of the study was to verify that the vascular responses were of neuronal origin and to examine the types of glutamate receptors involved in mediating the responses. We stimulated at 1 Hz for 8 s, with intervals of 30 s. Averages of only five stimulation periods were used in this protocol because of the transient effect of one of the drugs (GYKI, see below). CBF responses recorded with the LDFRed(250) were used for this part of the study.
Drug effects were evaluated as follows. The AMPA receptor antagonist 1-(4-aminophenyl)-4-methyl-7,8-methylene-dioxy-5H-2,3-benzodiazepine (GYKI 52466, Sigma) was dissolved in a concentration of 10 mg kg−1 (maximum, 2.5 mg ml−1), heated to 39 °C and slowly applied i.v. after control recordings had been obtained. The inhibitory effect of GYKI was calculated at 3-5 min after application when the effect was maximal. The effect of GYKI was fully reversed after 20 min (n= 6 rats). The AMPA receptor antagonist 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX, Research Biochemicals International) was dissolved in aCSF and applied topically at a concentration of 500 μm. The effect of CNQX was calculated at 10-15 min after application. The drug was washed out for the following 50 min and measurements were repeated (n= 6 rats). The selective and non-competitive NMDA receptor antagonist (5R,10S)-(5-methyl-10,11-dihydro-5H-dibenzo[a,d]cyclohepten-5,10-imine/dizocilpine (MK801, Sigma) was dissolved in saline and applied topically at a concentration of 10 μm. The effect of MK801 was tested at 50 min after application (n= 6 rats). In four rats MK801 (10 μm and 1 mm) was applied for 40 min, directly followed by topical application of CNQX (500 μm) for 35 min, and the combined effect was evaluated. Solutions were ultrasonicated when necessary. Drugs for topical application were dissolved in aCSF and kept at 37 °C while bubbling with 95 % air-5 % CO2, and pH was 7.30-7.40.