- Email: [email protected]
Capabilities of the WinLTP data acquisition program extending beyond basic LTP experimental functions
Journal of Neuroscience Methods 162 (2007) 346–356
Capabilities of the WinLTP data acquisition program extending beyond basic LTP experimental functions William W. Anderson ∗ , Graham L. Collingridge MRC Centre for Synaptic Plasticity and Department of Anatomy, University of Bristol, University Walk, Bristol BS8 1TD, UK Received 31 October 2006; received in revised form 21 December 2006; accepted 22 December 2006
Abstract WinLTP is a Windows data acquisition program designed for the investigation of long-term potentiation (LTP), long-term depression (LTD), and synaptic responses in general. The capabilities required for basic LTP and LTD experiments include alternating two-input extracellular pathway stimulation, LTP induction by single train, theta burst, and primed burst stimulation, and LTD induction by low frequency stimulation. WinLTP provides on-line analyses of synaptic waveforms including measurement of slope, peak amplitude, population-spike amplitude, average amplitude, area, rise time, decay time, duration, cell input resistance, and series resistance. WinLTP also has many advanced capabilities that extend beyond basic LTP experimental capabilities: (1) analysis of all the evoked synaptic potentials individually within a sweep, and the analysis of the entire train-evoked synaptic response as a single entity. (2) Multitasking can be used to run a Continuous Acquisition task (saving data to a gap-free Axon Binary File), while concurrently running the Stimulation/Acquisition Sweeps task. (3) Dynamic Protocol Scripting can be used to make more complicated protocols involving nested Loops (with counters), Delays, Sweeps (with various stimulations), and Run functions (which execute a block of functions). Protocol flow can be changed while the experiment is running. WinLTP runs on National Instruments M-Series and Molecular Devices Digidata 132x boards, and is available at www.winltp.com. © 2007 Elsevier B.V. All rights reserved. Keywords: Long-term potentiation; Long-term depression (LTD); Scripting; Multitasking; Epileptiform bursting; Kindling; Ischemia
1. Introduction Two experimental models commonly used to study the synaptic basic of learning and memory are long-term potentiation (LTP) and long-term depression (LTD). An important consideration for the electrophysiological study of LTP and LTD is data acquisition, on-line analysis, and the ability to generate the necessary stimulation protocols. A suitable program should be able to generate basic stimulation protocols including slow single pathway extracellular stimulation, or two pathway alternating stimulation. It should also be able to induce LTP by single train (Bliss and Lomo, 1973), theta burst (Larson et al., 1986), and primed burst stimulation (Rose and Dunwiddie, 1986), and to induce LTD by low frequency stimulation (Barrionuevo et al., 1980). This program should also be able to measure the peak amplitude of intracellular synaptic responses, and the slope and population-spike amplitude of extracellular synaptic responses ∗
Corresponding author. Tel.: +44 117 954 6573; fax: +44 117 929 1687. E-mail address: [email protected] (W.W. Anderson).
0165-0270/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.jneumeth.2006.12.018
(Bliss and Lomo, 1973). Finally, this program should also be able to analyze these responses on-line, to not only monitor the progress of the experiment, but also to alter amplitude and slope baselines during the experiment if necessary. See Bortolotto et al. (2001) for LTP experimental methods. Many data acquisition programs now perform several of these basic LTP experimental functions. Programs that do not require programming and are ‘ready to go’ include: Molecular Device’s pClamp (www.moldev.com), Heka’s PatchMaster (www.heka.com), AxoGraph (www.axographx.com), Cambridge Electronic Design’s Signal (www.ced.co.uk), NClamp (www.physiol.ucl.ac.uk/research/silver a/nclamp), Theta Burst’s NAC Gather (www.thetaburst.com), the Strathclyde Electrophysiology Software program WinWCP, and WinLTP’s DOS predecessor, the LTP Program (Anderson and Collingridge, 2001; www.ltp-program.com). Alternatively, many custom, in-house programs to perform LTP experiments have been written with Wave-Metrics’ Igor (www.wavemetrics.com), and National Instruments’ LabView (www.ni.com).
W.W. Anderson, G.L. Collingridge / Journal of Neuroscience Methods 162 (2007) 346–356
We have developed a ‘ready to go’ Windows program called WinLTP that is a successor to the DOS LTP Program, and performs basic acquisition, on-line analysis and protocol stimulation for studying LTP, LTD, and stimulus-evoked synaptic responses in general. Furthermore, one of the goals of writing WinLTP was to increase the complexity of LTP/LTD protocol stimulations that could easily be implemented. Therefore, WinLTP contains several advanced program functions that can further the study of LTP and a variety of other synaptic events. First, WinLTP can analyze all evoked synaptic potentials in a sweep, and also has special train analyses to analyze either every synaptic potential in a train, or analyze the train-evoked synaptic response as a whole. To our knowledge, the only other electrophysiological program with these capabilities is WinLTP’s predecessor, the LTP Program. Second, WinLTP is now a multitasking program that can run two independent tasks simultaneously: the repetitive Stimulation/Acquisition Sweeps task (similar to the LTP Program), and the Continuous Acquisition recording task. Third, the Stimulation/Acquisition Sweeps task now has Dynamic Protocol Scripting to produce complex protocols, changeable at run time. To our knowledge, PatchMaster is the only electrophysiological data acquisition program that has protocol scripting similar to WinLTP. In addition to LTP studies, WinLTP is also a useful tool for investigating ischemia, epilepsy, and synaptic responses in general. WinLTP has been available at www.winltp.com since September 2005, and has been presented in abstract form (Anderson and Collingridge, 2005). WinLTP’s predecessor, the LTP Program, has been used in over 200 publications. 2. Methods WinLTP is a Windows program that runs on computers using Windows 2000 or XP. It was written with Borland C++ Builder using Win32 VCL components. WinLTP uses National Instruments M-Series boards and Molecular Devices Digidata 1320A and 1322A boards. Other programs that use National Instruments M-Series boards include WinWCP, WinEDR, Nclamp, and custom in-house software using Igor and LabView. Other programs that use the Digidata 132x boards include pClamp, AxoScope, AxoGraph, WinWCP and WinEDR. The extracellular pathway S0 and S1 stimulation outputs go to digital outputs 0 and 1 of the data acquisition interface boxes and these trigger stimulus isolation units, and the analog IC0 output goes to the analog output 0 of the interface boxes and usually controls the command voltage of a patch clamp. 3. Results WinLTP (version 0.94) has extracellular and intracellular stimulation capabilities, records in extracellular, current clamp, voltage-clamp modes, and has many different on-line analyses of synaptic responses. The program produces repetitive sweeps with simultaneous data acquisition (up to 1,000,000 samples/sweep in two channels at up to 40 kHz sampling rate), extracellular pathway stimulation (in pulses or trains), and intracellular epoch-like analog stimulation (in mV or pA). WinLTP,
347
like the DOS LTP Program, has several basic protocols that facilitate running basic LTP and LTD experiments, including slow alternating dual extracellular pathway stimulation, LTP induction by single train, theta burst, and primed burst stimulation, and LTD induction by low frequency stimulation. WinLTP analyses include peak amplitude, slope, and population-spike amplitude, and measuring cell input resistance and patch electrode series resistance continuously throughout the experiment. However, in contrast to the most programs that perform basic LTP experiments, WinLTP also has several advanced capabilities that can extend the study of LTP and related synaptic events, which will be described later. 3.1. Basic LTP/LTD experiment Fig. 1 shows the layout of the WinLTP program to run a basic LTP/LTD experiment showing Protocol and Detection fields, Sweep Analysis graphs, Sweep Acquisition graphs, Stimulation fields and graphs, and the Spreadsheet and Run Buttons. In WinLTP, alternating dual pathway stimulation (S0 then S1) of the experiment in Fig. 1 is achieved by producing dual alternating sweeps (Pulse Sweep P0 then Pulse Sweep P1) in which Pulse Sweep P0 has one extracellular pathway stimulation, S0, and Pulse Sweep P1 has one extracellular pathway stimulation, S1. The MainProtocol panel shows this protocol involving alternating P0sweep (with one S0 stimulation pulse), then P1sweep (with one S1 stimulation pulse), in an AvgLoop to produce an alternating dual extracellular pathway stimulation with signal averaging every four sweeps. The Sweep Acquisition panel shows an S0-evoked fEPSP (averaged from four sweeps) with the slope marked by a red line. The Slope0 graph shows slope calculations for S0-evoked fEPSPs (red triangles) and S1-evoked fEPSPs (magenta squares). LTP S0 stimulation is evoked by clicking the ‘Single T0’ Run Button to produce a T0sweep containing 100 S0 pulses at 100 Hz. LTD S0 stimulation (900 S0 pulses, 1 s−1 ) is initiated by clicking the ‘Repeat P0’ Run Button to run 900 P0sweeps at 1 s−1 , and LTD S1 stimulation (900 S1 pulses, 1 s−1 ) is initiated by clicking the ‘Repeat P1’ Run Button to run 900 P1sweeps at 1 s−1 . WinLTP is currently capable of generating four different sweep stimulations with different stimulation capabilities on each. Two sweep stimulations are Pulse Sweeps P0 and P1, and these are often used for single pulse stimulation, can be repeated at set time intervals, and the sweep data can be signal averaged. The other two sweep stimulations are Train Sweeps T0 and T1, and these often evoke single, non-repetitive sweeps that contain train stimulation. WinLTP program output consists of one ASCII data file for each acquired sweep, spreadsheet analysis data saved to an ASCII file, and continuous acquisition data saved to a gap-free Axon Binary File (ABF). Each ASCII acquisition sweep file has a 64 line header that contains complete acquisition and stimulation information for that sweep (not shown). The ASCII sweep files can be reanalyzed with the off-line reanalysis version of WinLTP after the experiment. Both the ASCII sweep and analysis files can be directly imported into spreadsheet programs such as Excel, analysis
348
W.W. Anderson, G.L. Collingridge / Journal of Neuroscience Methods 162 (2007) 346–356
Fig. 1. WinLTP layout for a basic LTP/LTD experiment showing the Protocol fields (upper left panel), Analysis graphs (in this case only one slope graph, top right panel), Sweep Acquisition (middle right panel), Sweep Stimulation fields and graphs (lower left and right panels), and the Spreadsheet and Run Buttons (bottom panels). Detection fields to change synaptic potential detection values are hidden behind the Protocol fields in this screenshot. The MainProtocol panel shows the alternating P0sweep every 30 s, then P1sweep every 30 s, in an AvgLoop of four to produce an average every four sweeps. In the spreadsheet, “Time of Day” shows the time the sweep began, “Time m:s” shows the time of the stimulus pulse from when analysis starts, “Sx” shows whether S0 or S1 stimulation was used to evoke the synaptic response, “Pul#” shows the number of the S0 or S1 pulse that evokes the synaptic response, and “Slope” shows the calculated slope of the evoked response.
programs such as Igor, Origin, AxoGraph and NeuroMatic, and graphing programs such as SigmaPlot. The gap-free ABF files can be analyzed by programs such as DataView (http://www.st-andrews.ac.uk/∼wjh/dataview), MiniAnalysis (www.synaptosoft.com), ClampFit (www.moldev.com), WinEDR and AxoGraph. Both these file formats, ASCII and ABF, are two of the most standard in electrophysiology. 3.2. Stimulation during the sweep In WinLTP, stimulation during the sweep is comprised of epochs, similar to that used in pClamp. Fields in the Sweep Stimulation area control the sweep stimulation and include: (1) Sweep Duration, (2) S0 and S1 extracellular stimulation (organized in terms of pulses and trains), (3) IntraCellular (IC) Analog Output stimulation organized as epochs (up to six sequential steps, StepA to StepF), and (4) four Digital Outputs that can also be produced during the IntraCellular (IC) epochs.
The extracellular, intracellular and digital stimulation in each sweep is controlled by the fields in the P0, P1, T0 or T1 tabsheet in the Field Sweep Stimulation panels (Fig. 2, left and bottom), and plotted in the Graph Sweep Stimulation area (Fig. 2, upper right). The P0, P1, T0 or T1 tabs determine which sweep stimulation is to be examined and changed. The Field Sweep Stimulation area is functionally coupled with the Graph Stimulation area, so that when the P0 tabsheet in the Field Sweep Stimulation area is clicked on, the P0 Sweep Stimulation graph comes up. The S0, S1 and IC0 tabs determine whether S0 extracellular stimulation, S1 extracellular stimulation, or IC0 intracellular analog and digital stimulation is to be changed. S0 and S1 stimulation can be set to either ‘Off’ or ‘On’ in a pop-up menu, and if On can be either ‘Pulses’ (Fig. 2, upper left) or ‘Trains’ (Fig. 2, lower left). IC0 stimulation can be set to either ‘Off’, ‘Amplitude’ On, or ‘Amplitude + Digital Out’ On in a pop-up menu. If IC0 analog output is On, the output for each epoch can be: ‘Off’, ‘Step’ or ‘RsRm Step’ which is the step where patch
W.W. Anderson, G.L. Collingridge / Journal of Neuroscience Methods 162 (2007) 346–356
349
Fig. 2. An example of P0sweep stimulation consisting of two S0 pulses (left side of top panel), two S1 trains (left bottom panel), and IntraCellular Analog Output channel 0 (IC0), Digital Sync output (see the ‘S’ in DO2), and Digital Out step output (see the ‘1’s in DO3 and DO4) (right bottom panel). In this voltage clamp example, stimulation is in mV.
electrode series resistance and cell input resistance are measured relative to the previous step (Fig. 2, lower right). In this version of WinLTP, stimulations such as primed burst stimulation cannot currently be generated by only one extracellular output, but requires one of the digital output pulses being OR’ed with the S0 or S1 train output using a hardware OR gate. 3.3. Signal averaging, stimulus artifact blanking and digital filtering In addition to acquiring and analyzing raw sweeps, WinLTP can also average these sweeps, remove the stimulus artifacts, and filter the sweeps using a Gaussian digital filter (Colquhoun and Sigworth, 1983) in a manner similar to the DOS LTP Program (Anderson and Collingridge, 2001). However, in WinLTP
the signal averaging occurs within an AvgLoop protocol statement, and the sweep to be averaged must be a Pulse (P0 or P1) Sweep (see Fig. 1). Calculations of slopes and peaks are made on the latest processed sweep. For example, if signal averaging, stimulus artifact blanking and digital filtering are used, then the averaged, blanked and filtered sweep is the one that will be analyzed (see Fig. 3). The removal of stimulus artifacts allow accurate determination of synaptic area and peak during a train stimulation (Fig. 4b). 3.4. Measurement of cell input resistance and patch electrode series resistance WinLTP can measure cell input resistance and patch electrode series resistance (both in M) and plot the results
Fig. 3. Analyses of prestimulus DC baseline and peak amplitude to two S0 stimulus pulses. Whether a peak is detected as positive or negative can be determined automatically (Auto), or forced to be positive or negative. Patch electrode series resistance (Rs) (from the capacitance current transient peak) and membrane or cell input resistance (Rm) (from the steady state value) were both also calculated automatically in M during the response to a RsRm voltage-clamp pulse, and do not require input fields. The synaptic analyses were performed on the dark blue trace digitally filtered at 500 Hz for their most accurate measurement, but Rs and Rm analyses were performed on the raw, unfiltered gray trace for their most accurate measurement.
350
W.W. Anderson, G.L. Collingridge / Journal of Neuroscience Methods 162 (2007) 346–356
Fig. 4. Special analysis of train-evoked synaptic responses. (a) Analysis of every EPSP using the baseline of the first pulse. (b) Analysis of the train as a whole while blanking the stimulus artifacts to enable accurate measurement of area and peak amplitude.
on-line throughout an experiment. During an experiment, accurate on-line series resistance measurement is very important, because even small changes will affect EPSC amplitude, and series resistance often increases gradually during the experiment. Membrane or cell input resistance (Rm) can be measured during intracellular sharp electrode recording and patch clamp recording in both current- and voltage-clamp modes. Rm detection and measurement occurs automatically, and is the difference between the pre-RsRm Step baseline value and the value during the latter part of the RsRm Step. Patch electrode series resistance (Rs) can only be measured during patch clamping in voltage-clamp mode. In this version of WinLTP, Rs is calculated from the peak of the capacitative transient, generated by the RsRmStep stimulation epoch. Since WinLTP has a maximum sampling rate of 40 kHz the peak capacitative transient can be captured fairly well. To accurately measure Rs from the peak of the capacitative transient, the pipette capacitance must have been cancelled out, and series resistance compensation must be off. Rs is measured as Rs =
Vpulse Ipeak
where Ipeak is the peak amplitude of the capacitative current transient and Vpulse is the size of the voltage-clamp pulse. Rm
is measured as Rm =
Vpulse − Rs Iss
where Iss is the amplitude of the current at steady state measured in the latter part of the RsRm Step, and Vpulse is the size of the voltage-clamp pulse (see Ogden and Stanfield, 1994). This method of measuring series resistance tends to overestimate the value due to the initial slow rise of the voltage-clamp pulse (10–20 s) and low-pass filtering, but it does allow the researcher to detect changes in series resistance. This method of measuring series resistance is similar to that used by AxoGraph, WinWCP and Signal, but differs from the method used by pClamp. In pClamp, the transient portion of the current response is fitted by a single exponential, and series resistance is calculated from the time constant of this single exponential decay, the amplitude of the voltage pulse, and the total charge under the capacitative transient. When digital filtering is used to more accurately measure values such as peak amplitude, WinLTP also allows Rs measurements to be made from the raw, unfiltered trace for more accurate measurement of the capacitative current transient (Fig. 3). 3.5. Analyses of all synaptic responses in a sweep Until now, we have been discussing capabilities of WinLTP that are involved in basic LTP/LTD experiments. Now we
W.W. Anderson, G.L. Collingridge / Journal of Neuroscience Methods 162 (2007) 346–356
will begin discussing advanced capabilities of WinLTP. One main difference between the analysis of synaptic responses by WinLTP versus other programs that study LTP such as pClamp, AxoGraph, Theta Burst’s NAC Gather and Signal, is that these other programs usually analyze one or a few synaptic responses using timing determined by cursors or epochs. In contrast, WinLTP, like the DOS LTP Program, can analyze all the synaptic responses produced by up to 1000 extracellular S0 and S1 stimulation pulses in a sweep using timing determined by the occurrence of these S0 and S1 pulses. WinLTP analyses include the prestimulus DC baseline, slope (Fig. 1), peak amplitude and latency (Fig. 3), populationspike amplitude and latency, area, average amplitude, rise time, decay time, duration and coastline (see also Anderson and Collingridge, 2001). DC baseline gives an indication of DC shifts in intracellular current- and voltage-clamp recordings. Coastline calculates the total amount of vertical deflection between two time points after the stimulus pulse and is measured in mV or pA. For example, an EPSP of 1 mV amplitude would have a coastline measurement of 2 mV. Duration and coastline are particularly useful for analyzing epileptic events. For the analyses of each S0- and S1-synaptic response in a sweep, input fields are used to set the times before the stimulus pulse occurrence for DC baseline, or after the pulse occurrence for most other analyses. For example, to measure peak amplitude in Fig. 3 the DC baseline was measured 15–2 ms before each S0 pulse, and the peak amplitude was measured between 8 and
351
25 ms after each S0 pulse, and are shown by the dotted and solid lines in the graph. The values of these calculations are plotted in calculation graphs, and in the spreadsheet lines. The fact that WinLTP analyzes synaptic responses related to the occurrence of extracellular S0 or S1 stimulation pulses means that all S0- and S1-synaptic potentials in a sweep in both AD channels can be analyzed. 3.6. Special analysis of train-evoked synaptic responses Another difference between the analysis of synaptic responses in WinLTP versus other programs, is the ability of WinLTP (and its predecessor the LTP Program), to specially analyze train-evoked synaptic responses as individual synaptic responses or as a whole. First, sometimes it is important to examine each postsynaptic response evoked by each stimulus pulse in a train, in which case the baseline and synaptic response of each pulse is analyzed (as with the two pulse train in Fig. 3). Second, the synaptic responses evoked by train stimulation can be treated as a whole train in a special manner by WinLTP. The synaptic responses to all stimulus pulses in a train may be analyzed relative to the baseline of the first pulse in the train. Fig. 4a shows the peak amplitude measurement of each fEPSP in a four-pulse train, made relative to the baseline of the first pulse. Note the four measurements in the spreadsheet, one for each pulse.
Fig. 5. Thread diagram showing WinLTP as a multitasking program that has two tasks—the Stimulation/Acquisition Sweeps task and the Continuous Acquisition task, and uses five to seven threads of execution to achieve this.
352
W.W. Anderson, G.L. Collingridge / Journal of Neuroscience Methods 162 (2007) 346–356
Fig. 6. Two tasks operating simultaneously—the Continuous Acquisition task (top middle panel), and the Stimulation/Acquisition Sweeps task to measure patch electrode series resistance (Rs) and cell input resistance (Rm) (other graph panels).
Third, trains can also be analyzed as a single entity. If the baseline and response of only the first train pulse is used, if all stimulus artifacts are removed, and if the time of measurement is set sufficiently long to encompass the whole train, then the synaptic response of the entire train will be measured (Fig. 4b). With this analysis, the peak amplitude of the largest EPSP in the train and the area of the synaptic response of the train can be measured. Note, only one measurement is in the spreadsheet, for the first pulse, e.g. the whole train. 3.7. Multitasking The second advanced capability of WinLTP is multitasking, the ability to run several independent electrophysiological tasks simultaneously. WinLTP can currently run two tasks simultaneously: the repetitive Stimulation/Acquisition Sweeps task, and Continuous Acquisition ‘tape recording’ task (Fig. 5). These two tasks are produced by five or more separate threads of execution in WinLTP: the Stim Sweep In Thread and the Continuous Out Thread to produce stimulation/acquisition sweeps, the Continuous Acquisition In Thread to produce continuous acquisition, the User Interface Thread to capture user input and output screen graphics, and the Digidata 132x or M-Series board threads.
3.8. Continuous Acquisition task The second task is a ‘tape recorder’ task that saves continuously acquired data to a gap-free Axon Binary File for off-line analysis of spontaneous events using other programs such as DataView, MiniAnalysis, Clampfit, WinEDR or AxoGraph. A simple but useful example of this multitasking is to continuously acquire data with the Continuous Acquisition task, and to periodically output a stimulation/acquisition sweep with an RsRm Step pulse and perform on-line analysis and plotting of patch electrode series resistance (Rs) and cell input resistance (Rm) to determine the quality of the electrode and the cell throughout the experiment (Fig. 6). However, because the Continuous Acquisition task is completely separate from the Stimulation/Acquisition Sweeps task, any stimulation/acquisition protocol can be run with it. 3.9. Dynamic Protocol Scripting The third way WinLTP provides advanced LTP experimental functions is Dynamic Protocol Scripting. WinLTP actually has two modes of protocol operation, the LTP Protocol mode, which produces alternating P0/P1sweeps without or with averaging and requires no script programming (Fig. 1), and the Dynamic Protocol Scripting mode (Fig. 7). Dynamic Protocol Scripting enables
W.W. Anderson, G.L. Collingridge / Journal of Neuroscience Methods 162 (2007) 346–356
353
Fig. 7. One of the simplest yet useful LTP protocol scripts: minimal LTP stimulation protocol using a Delay line. The panel on the left shows the protocol script toolbox consisting of the Insert Buttons, the protocol script panel, and the Delete panel. The Delay value was changed from 180 to 60 s at arrow, but it only changed the period of the next delay.
more complicated protocols to be made involving loops, delays, and sweeps (with various stimulations). Statements include Run, Loop, AvgLoop, Delay and Sweep. The Dynamic Protocol Scripting is graphical in orientation with (1) a protocol script toolbox of Insert Buttons (for Run, AvgLoop, Loop, Delay and Sweeps), (2) a protocol script area showing lines of loops, sweeps and delays which set the protocol flow of execution, and (3) a Delete panel (Fig. 7). The Run, Loop, AvgLoop, Delay and Sweep lines can be pulled down from green Insert Buttons and dropped into the script, and they can be deleted by dragging them to the Delete panel. Because WinLTP scripting is based on a double-linked list instead of reading a static text based script, WinLTP protocol flow of execution can be rapidly changed while the experiment is running. This is easily done by checking/un-checking the check boxes beside the Run, Loop, AvgLoop, Delay and Sweep lines, and by changing the number of loops. Sweep Period and Delay Period times, and all sweep stimulation values (Fig. 2) can also be changed while the protocol is running. If the check box on the left of the Run, Loop, AvgLoop, Delay or Sweep statement is checked, then the Run, Loop, AvgLoop, Delay or Sweep will be run; otherwise it will not. Furthermore, by un-checking the check box on the Loop or AvgLoop, the loop will run to the bottom and exit (see Fig. 8a). Un-checking the check box for the Run, Delay and Sweep will not cause them to be prematurely terminated but they will not be run the next time protocol execution reaches them. The edit field on the right of the Loop and AvgLoop statements sets the number of loops. The edit field on the right of the Sweep and Delay statements determines the Sweep Period or Delay Period in seconds. Figs. 7 and 8 show examples of dynamic protocol scripts directly useful in LTP/LTD experiments. In all of these examples AD0 records S0 output and AD1 records S1 output (except in Fig. 8c where AD1 records the intracellular command voltage). P0sweep here has S0 single pulse stimulation, P1sweep here has S1 single pulse stimulation, T0sweep here usually has S0 train stimulation, and T1sweep here usually has S1 train stimulation, unless otherwise mentioned. Fig. 7 shows one of the simplest but useful script protocol, for minimal LTP stimulation, which consists of four averaged
pulse responses with a large delay in between (as in Frey et al., 1988). This is achieved simply adding a Delay below the AvgLoop containing the Sweep. Fig. 8a is an example of maintaining the stimulation frequency of one pathway (S0) while reducing the frequency of another pathway (S1). This is useful in LTP studies to investigate whether the rate of the slow repeat stimulation has an effect. When the check box of the middle continuous loop (“Loop [99999]”) is checked, the lower frequency stimulation of one pathway (S1) begins (at one-fourth the S0 frequency, Loop = 3). Un-checking the check box of the middle loop (down arrow) stops the lower frequency S1 stimulation. This code is an example of how checking and un-checking a continuous Loop can be used to start and stop different stimulations, and therefore change the protocol flow of execution. Fig. 8b shows a complex LTP/LTD induction protocol. An LTP inducing tetanus was delivered to one pathway (S0) immediately followed by an LTD induction by low frequency stimulation in the other pathway (S1) while the S0 pathway was returned to slow repeat stimulation. The induction stimulation begins when the check box on the Run line is checked. This produces a 1 s, 100 Hz LTP induction train in the S0 pathway by T0sweep, followed by rapid 1 Hz LTD stimulation in the second pathway (S1) by P1sweep, but with continuing S0 low frequency (1/30 s) stimulation from the T1sweep. The P0sweep contains one S0 pulse, the P1sweep contains one S1 pulse, the T0sweep contains a 2 s S0 train, and the T1sweep contains both a single S0 pulse and a single S1 pulse. Note that this LTP/LTD induction protocol cannot merely be produced by evoking LTP stimulation and then LTD stimulation because the slow repeat S0 pulse output continues during the more rapid S1 LTD stimulation output. Fig. 8c shows an example of LTP/LTD stimulation with steady depolarization. If the sweeps have a steady depolarization throughout the sweep, and the Sweep Duration is equal to the Sweep Period (in other words, there is no time between sweeps), then steady intracellular voltage changes can accompany the extracellular stimulation. The LTP/LTD stimulation here shows such steady depolarization when repetitive 0.5 Hz T0sweeps depolarized by 70 mV to produce LTP, and was immediately followed by repetitive 1 Hz T1sweeps depolarized by 40 mV to
354
W.W. Anderson, G.L. Collingridge / Journal of Neuroscience Methods 162 (2007) 346–356
Fig. 8. More LTP/LTD script protocols. (a) Lower stimulation frequency of one pathway (S1) vs. another pathway (S0). The lower frequency S1 stimulation (at one-fourth the S0 frequency, Loop = 3) was started by checking the inner continuous loop (“Loop [99999]”) at the up arrow and stopped by un-checking the Loop at the down arrow. (b) LTP train stimulation in the S0 pathway followed by LTD stimulation in the S1 pathway accompanied by continued slow repeat stimulation of the S0 pathway. Stimulation was started by checking the check box on the Run line (up arrow), and then un-checking it once the LTP/LTD stimulation had started (down arrow). (c) LTP/LTD stimulation with associated steady depolarization. LTP/LTD stimulation was started by clicking the Run check box (up arrow) and then un-checking it when LTP/LTD stimulation had started (down arrow). S0 pulse stimulation is recorded in AD0 and intracellular stimulation voltage from AnalogOut0 is recorded in AD1. P0, T0 and T1 sweeps here have S0 single pulse stimulation, T0sweep has steady 70 mV depolarization, and T1sweep has steady 40 mV depolarization.
produce LTD. This stimulation was started by checking on the check box on the Run line (up arrow), and then un-checking the Run line check box once the LTP/LTD stimulation has started (down arrow).
4. Discussion A versatile LTP program needs to perform several functions including special LTP analyses such as population-spike
W.W. Anderson, G.L. Collingridge / Journal of Neuroscience Methods 162 (2007) 346–356
amplitude, on-line analyses, alternating dual extracellular pathway stimulation, LTP induction by single train, theta burst and primed burst stimulation, and LTD induction by low frequency stimulation. These capabilities are now performed by many electrophysiological programs including pClamp, PatchMaster, AxoGraph, Signal, Nclamp, NAC Gather, WinWCP, and WinLTP’s DOS predecessor, the LTP Program. The present program, WinLTP, has several additional capabilities that makes it particularly useful for the study of LTP and LTD. 4.1. Advanced capabilities of WinLTP First, most electrophysiological data acquisition programs use cursors or stimulation epochs to determine the regions over which synaptic potentials will be analyzed. Therefore, only one or a few synaptic responses are usually analyzed in an acquisition sweep. In contrast, WinLTP, and its predecessor the LTP Program, can analyze up to 1000 S0- and S1-evoked synaptic potentials in a sweep, and in both AD channels. Furthermore, we have extended the analysis of synaptic potentials to those evoked by entire trains of stimulation. The synaptic responses evoked by train stimulation can either be analyzed as many individual synaptic responses, or the train analyzed as a whole, or each individual synaptic response can be measured relative to the baseline of the first pulse. When analyzing the synaptic responses to high frequency trains, we also found that contamination by stimulus artifacts prohibited accurate measurement of peaks and areas, so we added a function to automatically remove the stimulus artifacts. To our knowledge, the only other electrophysiological program with these capabilities is the LTP Program. Second, WinLTP is a multitasking program. Currently it simultaneously runs two independent tasks: the Stimulation/Acquisition Sweeps task, and a Continuous Acquisition recording task. This allows one, for example, to record miniature synaptic potentials or other spontaneously occurring events while at the same time to periodically analyzing and plotting on-line series resistance and cell input resistance. However, it is important to stress that because the tasks are independent, any stimulation/acquisition sweep protocol can be run with the Continuous Acquisition task. Furthermore, while the Continuous Acquisition task is running, the protocol flow, the sweep stimulation values, and the detection values in the Stimulation/Acquisition Sweeps task can still be changed. In other electrophysiological programs, such as pClamp, CED Spike2/Signal, WinEDR/WinWCP, NClamp and AxoGraph, the stimulation/acquisition sweep and continuous acquisition functions can be run separately, but not at the same time. All the standard electrophysiological software is multitasking in the sense that they can generate stimuli while simultaneously acquiring data during a sweep. WinEDR can even generate repetitive pulse or train stimulation during continuous acquisition. However, WinLTP is fundamentally different in that the two tasks can be completely independent of each other (in fact they are run by completely separate threads of execution). It is almost as if two separate programs are simultaneously running but are sharing one data acquisition board.
355
A preliminary non-released version of WinLTP also runs a third task, Capturing Spontaneous Events, which detects and plots spontaneously occurring events such spontaneous epileptiform bursts. However, because this is not fully implemented, it is not available in the current distributed version (0.94). Third, WinLTP has Dynamic Protocol Scripting which enables very complex protocols consisting of Sweeps, Delays, Run lines, and crucially, Loops with counters and loop within a loop constructs to be run. Dynamic Protocol Scripting was designed to be an easy to use, more general method of producing complicated protocols. Furthermore, because WinLTP scripting uses a double linked list database, the protocol flow of execution can be easily changed at runtime, as can the numbers of loops, Sweep and Delay Periods, and all sweep stimulations. To our knowledge, only PatchMaster has a similar method of protocol scripting that determines the protocol flow of execution and which includes loops with counters, and loop within a loop constructs, and can be substantially changed at runtime. A second type of protocol generation, using interpreted or compiled static text script, is performed by Signal. While these protocols can contain loop counters and loop within a loop constructs, they use static text script, and therefore the protocol flow of execution cannot be as easily changed at run time. A third type of protocol generation, using sweep or protocol linking which enables one sweep to link to the next, is performed by pClamp using Sequencing Keys, Signal using Protocol Dialog Boxes, AxoGraph, WinWCP and NAC Gather. These programs can form loops of sweeps, but because they have no loop counter (beyond repeating a single sweep a given number of times), they therefore have no automatic way of exiting the loop (except by user input). They therefore have no loop within a loop construct, and do not have a general method of producing complicated protocols. 4.2. Wider uses for WinLTP in studying synaptic physiology Although WinLTP was particularly developed for studying LTP, its flexible protocol generation, sweep stimulation, on-line analyses and multitasking continuous acquisition make it useful for studying a wide range of synaptic phenomena including ischemia and epilepsy. WinLTP, and its predecessor the LTP Program, are being used in studies of ischemia (Coelho et al., 2006). WinLTP also has several capabilities useful for studying epilepsy. WinLTP can train evoke and record electrographic seizures in 100 s long sweeps. WinLTP cannot only induce electrographic seizures with normal repetitive trains (as in Lothman and Williamson, 1993), it can also apply more complex induction stimulation protocols such as placing single trigger pulses between the induction trains to assess the excitability of epileptiform bursts (as in Anderson et al., 1987). And finally, WinLTP can measure the duration and coastline of stimulus-evoked epileptiform bursts and electrographic seizures, measurements that are particularly useful in epilepsy studies. 4.3. Future additions to WinLTP The multitasking and Dynamic Protocol Scripting coupled with the sweep stimulation/acquisition and special LTP analysis
356
W.W. Anderson, G.L. Collingridge / Journal of Neuroscience Methods 162 (2007) 346–356
capabilities provide a solid foundation for basic and advanced data acquisition techniques with which to study LTP and related synaptic events. However, several capabilities must be added to complete WinLTP’s advanced experimental capabilities. Protocol scripting should include if statements to test protocol flow depending on whether amplitude/slope baseline stability has been achieved. Improved sweep stimulation will include five extracellular stimulation outputs (S0–S4), extracellular stimulation with many sequential trains to directly perform primed burst stimulation, intracellular stimulation with trains and ramps, and increment/decrement of stimulation values such as varying intracellular analog voltage and interpulse intervals. The number of Pulse and Train Sweeps will also be increased to five (P0–P4 and T0–T4). Better analysis of patch electrode series resistance (Rs) will require exponential curve fitting of the capacitative current transient. Most innovative will be implementation of the capture and analysis of Spontaneous Events task to study events such as spontaneous epileptiform bursts and electrographic seizures. Acknowledgements We would like to thank Steve Fitzjohn for reading the manuscript, Neil Bannister for discussions, and Zuner Bortolotto, Stephane Peineau and Tom Sanderson for data in the figures. Supported by the MRC. References Anderson WW, Collingridge GL. The LTP Program: a data acquisition program for on-line analysis of long-term potentiation and other synaptic events. J Neurosci Methods 2001;108:71–83. Anderson WW, Collingridge GL. WinLTP: a Windows data acquisition program for on-line analysis of long-term potentiation and other synaptic events,
Program No. 569.19. 2005, Abstract Viewer/Itinerary Planner. Washington, DC: Society for Neuroscience; 2005. Online. Anderson WW, Swartzwelder HS, Wilson WA. The NMDA receptor antagonist 2-amino-5-phosphonovalerate blocks stimulus train-induced epileptogenesis but not epileptiform bursting in the rat hippocampal slice. J Neurophysiol 1987;57:1–21. Barrionuevo G, Schottler F, Lynch G. The effects of repetitive slow-frequency stimulation on control and ‘potentiated’ synaptic responses in the hippocampus. Life Sci 1980;27:2385–91. Bliss TVP, Lomo T. Long-lasting potentiation of synaptic transmission in the dentate area of the anaesthetized rabbit following stimulation of the perforant path. J Physiol 1973;232:331–56. Bortolotto ZA, Anderson WW, Isaac JTR, Collingridge GL. Synaptic plasticity in the hippocampal slice preparation. In: Crawley JN, Gerfen CR, Rogawski MA, Sibley DR, Skolnick P, Wray S, editors. Current protocols in neuroscience. New York: John Wiley & Sons; 2001. p. 6.13.1–23. Colquhoun D, Sigworth FJ. Fitting and statistical analysis of single channel records. In: Sakmann B, Neher E, editors. Single Channel Recording. London: Plenum Press; 1983. pp. 191–263. Coelho JE, Rebola N, Fragata I, Ribeiro JA, De Mendonca A, Cunha RA. Hypoxia-induced desensitization and internalization of adenosine A(1) receptors in the rat hippocampus. Neuroscience 2006;138:1195– 203. Frey U, Krug M, Reymann KG, Matthies H. Anisomycin, an inhibitor of proteinsynthesis, blocks late phases of LTP phenomena in the hippocampal CA1 region in vitro. Brain Res 1988;452:57–65. Larson J, Wong D, Lynch G. Patterned stimulation at the theta frequency is optimal for the induction of hippocampal long-term potentiation. Brain Res 1986;368:347–50. Lothman EW, Williamson JM. Rapid kindling with recurrent hippocampal seizures. Epilepsy Res 1993;14:209–20. Ogden D, Stanfield P. Patch clamp techniques for single channel and whole-cell recording. In: Ogden D, editor. Microelectrode techniques, the plymouth workshop handbook. 2nd ed. Cambridge: The Company of Biologists Ltd.; 1994. p. 53–78. Rose GM, Dunwiddie TV. Induction of hippocampal long-term potentiation using physiologically patterned stimulation. Neurosci Lett 1986;69:244– 8.
Capabilities of the WinLTP data acquisition program extending beyond basic LTP experimental functions William W. Anderson ∗ , Graham L. Collingridge MRC Centre for Synaptic Plasticity and Department of Anatomy, University of Bristol, University Walk, Bristol BS8 1TD, UK Received 31 October 2006; received in revised form 21 December 2006; accepted 22 December 2006
Abstract WinLTP is a Windows data acquisition program designed for the investigation of long-term potentiation (LTP), long-term depression (LTD), and synaptic responses in general. The capabilities required for basic LTP and LTD experiments include alternating two-input extracellular pathway stimulation, LTP induction by single train, theta burst, and primed burst stimulation, and LTD induction by low frequency stimulation. WinLTP provides on-line analyses of synaptic waveforms including measurement of slope, peak amplitude, population-spike amplitude, average amplitude, area, rise time, decay time, duration, cell input resistance, and series resistance. WinLTP also has many advanced capabilities that extend beyond basic LTP experimental capabilities: (1) analysis of all the evoked synaptic potentials individually within a sweep, and the analysis of the entire train-evoked synaptic response as a single entity. (2) Multitasking can be used to run a Continuous Acquisition task (saving data to a gap-free Axon Binary File), while concurrently running the Stimulation/Acquisition Sweeps task. (3) Dynamic Protocol Scripting can be used to make more complicated protocols involving nested Loops (with counters), Delays, Sweeps (with various stimulations), and Run functions (which execute a block of functions). Protocol flow can be changed while the experiment is running. WinLTP runs on National Instruments M-Series and Molecular Devices Digidata 132x boards, and is available at www.winltp.com. © 2007 Elsevier B.V. All rights reserved. Keywords: Long-term potentiation; Long-term depression (LTD); Scripting; Multitasking; Epileptiform bursting; Kindling; Ischemia
1. Introduction Two experimental models commonly used to study the synaptic basic of learning and memory are long-term potentiation (LTP) and long-term depression (LTD). An important consideration for the electrophysiological study of LTP and LTD is data acquisition, on-line analysis, and the ability to generate the necessary stimulation protocols. A suitable program should be able to generate basic stimulation protocols including slow single pathway extracellular stimulation, or two pathway alternating stimulation. It should also be able to induce LTP by single train (Bliss and Lomo, 1973), theta burst (Larson et al., 1986), and primed burst stimulation (Rose and Dunwiddie, 1986), and to induce LTD by low frequency stimulation (Barrionuevo et al., 1980). This program should also be able to measure the peak amplitude of intracellular synaptic responses, and the slope and population-spike amplitude of extracellular synaptic responses ∗
Corresponding author. Tel.: +44 117 954 6573; fax: +44 117 929 1687. E-mail address: [email protected] (W.W. Anderson).
0165-0270/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.jneumeth.2006.12.018
(Bliss and Lomo, 1973). Finally, this program should also be able to analyze these responses on-line, to not only monitor the progress of the experiment, but also to alter amplitude and slope baselines during the experiment if necessary. See Bortolotto et al. (2001) for LTP experimental methods. Many data acquisition programs now perform several of these basic LTP experimental functions. Programs that do not require programming and are ‘ready to go’ include: Molecular Device’s pClamp (www.moldev.com), Heka’s PatchMaster (www.heka.com), AxoGraph (www.axographx.com), Cambridge Electronic Design’s Signal (www.ced.co.uk), NClamp (www.physiol.ucl.ac.uk/research/silver a/nclamp), Theta Burst’s NAC Gather (www.thetaburst.com), the Strathclyde Electrophysiology Software program WinWCP, and WinLTP’s DOS predecessor, the LTP Program (Anderson and Collingridge, 2001; www.ltp-program.com). Alternatively, many custom, in-house programs to perform LTP experiments have been written with Wave-Metrics’ Igor (www.wavemetrics.com), and National Instruments’ LabView (www.ni.com).
W.W. Anderson, G.L. Collingridge / Journal of Neuroscience Methods 162 (2007) 346–356
We have developed a ‘ready to go’ Windows program called WinLTP that is a successor to the DOS LTP Program, and performs basic acquisition, on-line analysis and protocol stimulation for studying LTP, LTD, and stimulus-evoked synaptic responses in general. Furthermore, one of the goals of writing WinLTP was to increase the complexity of LTP/LTD protocol stimulations that could easily be implemented. Therefore, WinLTP contains several advanced program functions that can further the study of LTP and a variety of other synaptic events. First, WinLTP can analyze all evoked synaptic potentials in a sweep, and also has special train analyses to analyze either every synaptic potential in a train, or analyze the train-evoked synaptic response as a whole. To our knowledge, the only other electrophysiological program with these capabilities is WinLTP’s predecessor, the LTP Program. Second, WinLTP is now a multitasking program that can run two independent tasks simultaneously: the repetitive Stimulation/Acquisition Sweeps task (similar to the LTP Program), and the Continuous Acquisition recording task. Third, the Stimulation/Acquisition Sweeps task now has Dynamic Protocol Scripting to produce complex protocols, changeable at run time. To our knowledge, PatchMaster is the only electrophysiological data acquisition program that has protocol scripting similar to WinLTP. In addition to LTP studies, WinLTP is also a useful tool for investigating ischemia, epilepsy, and synaptic responses in general. WinLTP has been available at www.winltp.com since September 2005, and has been presented in abstract form (Anderson and Collingridge, 2005). WinLTP’s predecessor, the LTP Program, has been used in over 200 publications. 2. Methods WinLTP is a Windows program that runs on computers using Windows 2000 or XP. It was written with Borland C++ Builder using Win32 VCL components. WinLTP uses National Instruments M-Series boards and Molecular Devices Digidata 1320A and 1322A boards. Other programs that use National Instruments M-Series boards include WinWCP, WinEDR, Nclamp, and custom in-house software using Igor and LabView. Other programs that use the Digidata 132x boards include pClamp, AxoScope, AxoGraph, WinWCP and WinEDR. The extracellular pathway S0 and S1 stimulation outputs go to digital outputs 0 and 1 of the data acquisition interface boxes and these trigger stimulus isolation units, and the analog IC0 output goes to the analog output 0 of the interface boxes and usually controls the command voltage of a patch clamp. 3. Results WinLTP (version 0.94) has extracellular and intracellular stimulation capabilities, records in extracellular, current clamp, voltage-clamp modes, and has many different on-line analyses of synaptic responses. The program produces repetitive sweeps with simultaneous data acquisition (up to 1,000,000 samples/sweep in two channels at up to 40 kHz sampling rate), extracellular pathway stimulation (in pulses or trains), and intracellular epoch-like analog stimulation (in mV or pA). WinLTP,
347
like the DOS LTP Program, has several basic protocols that facilitate running basic LTP and LTD experiments, including slow alternating dual extracellular pathway stimulation, LTP induction by single train, theta burst, and primed burst stimulation, and LTD induction by low frequency stimulation. WinLTP analyses include peak amplitude, slope, and population-spike amplitude, and measuring cell input resistance and patch electrode series resistance continuously throughout the experiment. However, in contrast to the most programs that perform basic LTP experiments, WinLTP also has several advanced capabilities that can extend the study of LTP and related synaptic events, which will be described later. 3.1. Basic LTP/LTD experiment Fig. 1 shows the layout of the WinLTP program to run a basic LTP/LTD experiment showing Protocol and Detection fields, Sweep Analysis graphs, Sweep Acquisition graphs, Stimulation fields and graphs, and the Spreadsheet and Run Buttons. In WinLTP, alternating dual pathway stimulation (S0 then S1) of the experiment in Fig. 1 is achieved by producing dual alternating sweeps (Pulse Sweep P0 then Pulse Sweep P1) in which Pulse Sweep P0 has one extracellular pathway stimulation, S0, and Pulse Sweep P1 has one extracellular pathway stimulation, S1. The MainProtocol panel shows this protocol involving alternating P0sweep (with one S0 stimulation pulse), then P1sweep (with one S1 stimulation pulse), in an AvgLoop to produce an alternating dual extracellular pathway stimulation with signal averaging every four sweeps. The Sweep Acquisition panel shows an S0-evoked fEPSP (averaged from four sweeps) with the slope marked by a red line. The Slope0 graph shows slope calculations for S0-evoked fEPSPs (red triangles) and S1-evoked fEPSPs (magenta squares). LTP S0 stimulation is evoked by clicking the ‘Single T0’ Run Button to produce a T0sweep containing 100 S0 pulses at 100 Hz. LTD S0 stimulation (900 S0 pulses, 1 s−1 ) is initiated by clicking the ‘Repeat P0’ Run Button to run 900 P0sweeps at 1 s−1 , and LTD S1 stimulation (900 S1 pulses, 1 s−1 ) is initiated by clicking the ‘Repeat P1’ Run Button to run 900 P1sweeps at 1 s−1 . WinLTP is currently capable of generating four different sweep stimulations with different stimulation capabilities on each. Two sweep stimulations are Pulse Sweeps P0 and P1, and these are often used for single pulse stimulation, can be repeated at set time intervals, and the sweep data can be signal averaged. The other two sweep stimulations are Train Sweeps T0 and T1, and these often evoke single, non-repetitive sweeps that contain train stimulation. WinLTP program output consists of one ASCII data file for each acquired sweep, spreadsheet analysis data saved to an ASCII file, and continuous acquisition data saved to a gap-free Axon Binary File (ABF). Each ASCII acquisition sweep file has a 64 line header that contains complete acquisition and stimulation information for that sweep (not shown). The ASCII sweep files can be reanalyzed with the off-line reanalysis version of WinLTP after the experiment. Both the ASCII sweep and analysis files can be directly imported into spreadsheet programs such as Excel, analysis
348
W.W. Anderson, G.L. Collingridge / Journal of Neuroscience Methods 162 (2007) 346–356
Fig. 1. WinLTP layout for a basic LTP/LTD experiment showing the Protocol fields (upper left panel), Analysis graphs (in this case only one slope graph, top right panel), Sweep Acquisition (middle right panel), Sweep Stimulation fields and graphs (lower left and right panels), and the Spreadsheet and Run Buttons (bottom panels). Detection fields to change synaptic potential detection values are hidden behind the Protocol fields in this screenshot. The MainProtocol panel shows the alternating P0sweep every 30 s, then P1sweep every 30 s, in an AvgLoop of four to produce an average every four sweeps. In the spreadsheet, “Time of Day” shows the time the sweep began, “Time m:s” shows the time of the stimulus pulse from when analysis starts, “Sx” shows whether S0 or S1 stimulation was used to evoke the synaptic response, “Pul#” shows the number of the S0 or S1 pulse that evokes the synaptic response, and “Slope” shows the calculated slope of the evoked response.
programs such as Igor, Origin, AxoGraph and NeuroMatic, and graphing programs such as SigmaPlot. The gap-free ABF files can be analyzed by programs such as DataView (http://www.st-andrews.ac.uk/∼wjh/dataview), MiniAnalysis (www.synaptosoft.com), ClampFit (www.moldev.com), WinEDR and AxoGraph. Both these file formats, ASCII and ABF, are two of the most standard in electrophysiology. 3.2. Stimulation during the sweep In WinLTP, stimulation during the sweep is comprised of epochs, similar to that used in pClamp. Fields in the Sweep Stimulation area control the sweep stimulation and include: (1) Sweep Duration, (2) S0 and S1 extracellular stimulation (organized in terms of pulses and trains), (3) IntraCellular (IC) Analog Output stimulation organized as epochs (up to six sequential steps, StepA to StepF), and (4) four Digital Outputs that can also be produced during the IntraCellular (IC) epochs.
The extracellular, intracellular and digital stimulation in each sweep is controlled by the fields in the P0, P1, T0 or T1 tabsheet in the Field Sweep Stimulation panels (Fig. 2, left and bottom), and plotted in the Graph Sweep Stimulation area (Fig. 2, upper right). The P0, P1, T0 or T1 tabs determine which sweep stimulation is to be examined and changed. The Field Sweep Stimulation area is functionally coupled with the Graph Stimulation area, so that when the P0 tabsheet in the Field Sweep Stimulation area is clicked on, the P0 Sweep Stimulation graph comes up. The S0, S1 and IC0 tabs determine whether S0 extracellular stimulation, S1 extracellular stimulation, or IC0 intracellular analog and digital stimulation is to be changed. S0 and S1 stimulation can be set to either ‘Off’ or ‘On’ in a pop-up menu, and if On can be either ‘Pulses’ (Fig. 2, upper left) or ‘Trains’ (Fig. 2, lower left). IC0 stimulation can be set to either ‘Off’, ‘Amplitude’ On, or ‘Amplitude + Digital Out’ On in a pop-up menu. If IC0 analog output is On, the output for each epoch can be: ‘Off’, ‘Step’ or ‘RsRm Step’ which is the step where patch
W.W. Anderson, G.L. Collingridge / Journal of Neuroscience Methods 162 (2007) 346–356
349
Fig. 2. An example of P0sweep stimulation consisting of two S0 pulses (left side of top panel), two S1 trains (left bottom panel), and IntraCellular Analog Output channel 0 (IC0), Digital Sync output (see the ‘S’ in DO2), and Digital Out step output (see the ‘1’s in DO3 and DO4) (right bottom panel). In this voltage clamp example, stimulation is in mV.
electrode series resistance and cell input resistance are measured relative to the previous step (Fig. 2, lower right). In this version of WinLTP, stimulations such as primed burst stimulation cannot currently be generated by only one extracellular output, but requires one of the digital output pulses being OR’ed with the S0 or S1 train output using a hardware OR gate. 3.3. Signal averaging, stimulus artifact blanking and digital filtering In addition to acquiring and analyzing raw sweeps, WinLTP can also average these sweeps, remove the stimulus artifacts, and filter the sweeps using a Gaussian digital filter (Colquhoun and Sigworth, 1983) in a manner similar to the DOS LTP Program (Anderson and Collingridge, 2001). However, in WinLTP
the signal averaging occurs within an AvgLoop protocol statement, and the sweep to be averaged must be a Pulse (P0 or P1) Sweep (see Fig. 1). Calculations of slopes and peaks are made on the latest processed sweep. For example, if signal averaging, stimulus artifact blanking and digital filtering are used, then the averaged, blanked and filtered sweep is the one that will be analyzed (see Fig. 3). The removal of stimulus artifacts allow accurate determination of synaptic area and peak during a train stimulation (Fig. 4b). 3.4. Measurement of cell input resistance and patch electrode series resistance WinLTP can measure cell input resistance and patch electrode series resistance (both in M) and plot the results
Fig. 3. Analyses of prestimulus DC baseline and peak amplitude to two S0 stimulus pulses. Whether a peak is detected as positive or negative can be determined automatically (Auto), or forced to be positive or negative. Patch electrode series resistance (Rs) (from the capacitance current transient peak) and membrane or cell input resistance (Rm) (from the steady state value) were both also calculated automatically in M during the response to a RsRm voltage-clamp pulse, and do not require input fields. The synaptic analyses were performed on the dark blue trace digitally filtered at 500 Hz for their most accurate measurement, but Rs and Rm analyses were performed on the raw, unfiltered gray trace for their most accurate measurement.
350
W.W. Anderson, G.L. Collingridge / Journal of Neuroscience Methods 162 (2007) 346–356
Fig. 4. Special analysis of train-evoked synaptic responses. (a) Analysis of every EPSP using the baseline of the first pulse. (b) Analysis of the train as a whole while blanking the stimulus artifacts to enable accurate measurement of area and peak amplitude.
on-line throughout an experiment. During an experiment, accurate on-line series resistance measurement is very important, because even small changes will affect EPSC amplitude, and series resistance often increases gradually during the experiment. Membrane or cell input resistance (Rm) can be measured during intracellular sharp electrode recording and patch clamp recording in both current- and voltage-clamp modes. Rm detection and measurement occurs automatically, and is the difference between the pre-RsRm Step baseline value and the value during the latter part of the RsRm Step. Patch electrode series resistance (Rs) can only be measured during patch clamping in voltage-clamp mode. In this version of WinLTP, Rs is calculated from the peak of the capacitative transient, generated by the RsRmStep stimulation epoch. Since WinLTP has a maximum sampling rate of 40 kHz the peak capacitative transient can be captured fairly well. To accurately measure Rs from the peak of the capacitative transient, the pipette capacitance must have been cancelled out, and series resistance compensation must be off. Rs is measured as Rs =
Vpulse Ipeak
where Ipeak is the peak amplitude of the capacitative current transient and Vpulse is the size of the voltage-clamp pulse. Rm
is measured as Rm =
Vpulse − Rs Iss
where Iss is the amplitude of the current at steady state measured in the latter part of the RsRm Step, and Vpulse is the size of the voltage-clamp pulse (see Ogden and Stanfield, 1994). This method of measuring series resistance tends to overestimate the value due to the initial slow rise of the voltage-clamp pulse (10–20 s) and low-pass filtering, but it does allow the researcher to detect changes in series resistance. This method of measuring series resistance is similar to that used by AxoGraph, WinWCP and Signal, but differs from the method used by pClamp. In pClamp, the transient portion of the current response is fitted by a single exponential, and series resistance is calculated from the time constant of this single exponential decay, the amplitude of the voltage pulse, and the total charge under the capacitative transient. When digital filtering is used to more accurately measure values such as peak amplitude, WinLTP also allows Rs measurements to be made from the raw, unfiltered trace for more accurate measurement of the capacitative current transient (Fig. 3). 3.5. Analyses of all synaptic responses in a sweep Until now, we have been discussing capabilities of WinLTP that are involved in basic LTP/LTD experiments. Now we
W.W. Anderson, G.L. Collingridge / Journal of Neuroscience Methods 162 (2007) 346–356
will begin discussing advanced capabilities of WinLTP. One main difference between the analysis of synaptic responses by WinLTP versus other programs that study LTP such as pClamp, AxoGraph, Theta Burst’s NAC Gather and Signal, is that these other programs usually analyze one or a few synaptic responses using timing determined by cursors or epochs. In contrast, WinLTP, like the DOS LTP Program, can analyze all the synaptic responses produced by up to 1000 extracellular S0 and S1 stimulation pulses in a sweep using timing determined by the occurrence of these S0 and S1 pulses. WinLTP analyses include the prestimulus DC baseline, slope (Fig. 1), peak amplitude and latency (Fig. 3), populationspike amplitude and latency, area, average amplitude, rise time, decay time, duration and coastline (see also Anderson and Collingridge, 2001). DC baseline gives an indication of DC shifts in intracellular current- and voltage-clamp recordings. Coastline calculates the total amount of vertical deflection between two time points after the stimulus pulse and is measured in mV or pA. For example, an EPSP of 1 mV amplitude would have a coastline measurement of 2 mV. Duration and coastline are particularly useful for analyzing epileptic events. For the analyses of each S0- and S1-synaptic response in a sweep, input fields are used to set the times before the stimulus pulse occurrence for DC baseline, or after the pulse occurrence for most other analyses. For example, to measure peak amplitude in Fig. 3 the DC baseline was measured 15–2 ms before each S0 pulse, and the peak amplitude was measured between 8 and
351
25 ms after each S0 pulse, and are shown by the dotted and solid lines in the graph. The values of these calculations are plotted in calculation graphs, and in the spreadsheet lines. The fact that WinLTP analyzes synaptic responses related to the occurrence of extracellular S0 or S1 stimulation pulses means that all S0- and S1-synaptic potentials in a sweep in both AD channels can be analyzed. 3.6. Special analysis of train-evoked synaptic responses Another difference between the analysis of synaptic responses in WinLTP versus other programs, is the ability of WinLTP (and its predecessor the LTP Program), to specially analyze train-evoked synaptic responses as individual synaptic responses or as a whole. First, sometimes it is important to examine each postsynaptic response evoked by each stimulus pulse in a train, in which case the baseline and synaptic response of each pulse is analyzed (as with the two pulse train in Fig. 3). Second, the synaptic responses evoked by train stimulation can be treated as a whole train in a special manner by WinLTP. The synaptic responses to all stimulus pulses in a train may be analyzed relative to the baseline of the first pulse in the train. Fig. 4a shows the peak amplitude measurement of each fEPSP in a four-pulse train, made relative to the baseline of the first pulse. Note the four measurements in the spreadsheet, one for each pulse.
Fig. 5. Thread diagram showing WinLTP as a multitasking program that has two tasks—the Stimulation/Acquisition Sweeps task and the Continuous Acquisition task, and uses five to seven threads of execution to achieve this.
352
W.W. Anderson, G.L. Collingridge / Journal of Neuroscience Methods 162 (2007) 346–356
Fig. 6. Two tasks operating simultaneously—the Continuous Acquisition task (top middle panel), and the Stimulation/Acquisition Sweeps task to measure patch electrode series resistance (Rs) and cell input resistance (Rm) (other graph panels).
Third, trains can also be analyzed as a single entity. If the baseline and response of only the first train pulse is used, if all stimulus artifacts are removed, and if the time of measurement is set sufficiently long to encompass the whole train, then the synaptic response of the entire train will be measured (Fig. 4b). With this analysis, the peak amplitude of the largest EPSP in the train and the area of the synaptic response of the train can be measured. Note, only one measurement is in the spreadsheet, for the first pulse, e.g. the whole train. 3.7. Multitasking The second advanced capability of WinLTP is multitasking, the ability to run several independent electrophysiological tasks simultaneously. WinLTP can currently run two tasks simultaneously: the repetitive Stimulation/Acquisition Sweeps task, and Continuous Acquisition ‘tape recording’ task (Fig. 5). These two tasks are produced by five or more separate threads of execution in WinLTP: the Stim Sweep In Thread and the Continuous Out Thread to produce stimulation/acquisition sweeps, the Continuous Acquisition In Thread to produce continuous acquisition, the User Interface Thread to capture user input and output screen graphics, and the Digidata 132x or M-Series board threads.
3.8. Continuous Acquisition task The second task is a ‘tape recorder’ task that saves continuously acquired data to a gap-free Axon Binary File for off-line analysis of spontaneous events using other programs such as DataView, MiniAnalysis, Clampfit, WinEDR or AxoGraph. A simple but useful example of this multitasking is to continuously acquire data with the Continuous Acquisition task, and to periodically output a stimulation/acquisition sweep with an RsRm Step pulse and perform on-line analysis and plotting of patch electrode series resistance (Rs) and cell input resistance (Rm) to determine the quality of the electrode and the cell throughout the experiment (Fig. 6). However, because the Continuous Acquisition task is completely separate from the Stimulation/Acquisition Sweeps task, any stimulation/acquisition protocol can be run with it. 3.9. Dynamic Protocol Scripting The third way WinLTP provides advanced LTP experimental functions is Dynamic Protocol Scripting. WinLTP actually has two modes of protocol operation, the LTP Protocol mode, which produces alternating P0/P1sweeps without or with averaging and requires no script programming (Fig. 1), and the Dynamic Protocol Scripting mode (Fig. 7). Dynamic Protocol Scripting enables
W.W. Anderson, G.L. Collingridge / Journal of Neuroscience Methods 162 (2007) 346–356
353
Fig. 7. One of the simplest yet useful LTP protocol scripts: minimal LTP stimulation protocol using a Delay line. The panel on the left shows the protocol script toolbox consisting of the Insert Buttons, the protocol script panel, and the Delete panel. The Delay value was changed from 180 to 60 s at arrow, but it only changed the period of the next delay.
more complicated protocols to be made involving loops, delays, and sweeps (with various stimulations). Statements include Run, Loop, AvgLoop, Delay and Sweep. The Dynamic Protocol Scripting is graphical in orientation with (1) a protocol script toolbox of Insert Buttons (for Run, AvgLoop, Loop, Delay and Sweeps), (2) a protocol script area showing lines of loops, sweeps and delays which set the protocol flow of execution, and (3) a Delete panel (Fig. 7). The Run, Loop, AvgLoop, Delay and Sweep lines can be pulled down from green Insert Buttons and dropped into the script, and they can be deleted by dragging them to the Delete panel. Because WinLTP scripting is based on a double-linked list instead of reading a static text based script, WinLTP protocol flow of execution can be rapidly changed while the experiment is running. This is easily done by checking/un-checking the check boxes beside the Run, Loop, AvgLoop, Delay and Sweep lines, and by changing the number of loops. Sweep Period and Delay Period times, and all sweep stimulation values (Fig. 2) can also be changed while the protocol is running. If the check box on the left of the Run, Loop, AvgLoop, Delay or Sweep statement is checked, then the Run, Loop, AvgLoop, Delay or Sweep will be run; otherwise it will not. Furthermore, by un-checking the check box on the Loop or AvgLoop, the loop will run to the bottom and exit (see Fig. 8a). Un-checking the check box for the Run, Delay and Sweep will not cause them to be prematurely terminated but they will not be run the next time protocol execution reaches them. The edit field on the right of the Loop and AvgLoop statements sets the number of loops. The edit field on the right of the Sweep and Delay statements determines the Sweep Period or Delay Period in seconds. Figs. 7 and 8 show examples of dynamic protocol scripts directly useful in LTP/LTD experiments. In all of these examples AD0 records S0 output and AD1 records S1 output (except in Fig. 8c where AD1 records the intracellular command voltage). P0sweep here has S0 single pulse stimulation, P1sweep here has S1 single pulse stimulation, T0sweep here usually has S0 train stimulation, and T1sweep here usually has S1 train stimulation, unless otherwise mentioned. Fig. 7 shows one of the simplest but useful script protocol, for minimal LTP stimulation, which consists of four averaged
pulse responses with a large delay in between (as in Frey et al., 1988). This is achieved simply adding a Delay below the AvgLoop containing the Sweep. Fig. 8a is an example of maintaining the stimulation frequency of one pathway (S0) while reducing the frequency of another pathway (S1). This is useful in LTP studies to investigate whether the rate of the slow repeat stimulation has an effect. When the check box of the middle continuous loop (“Loop [99999]”) is checked, the lower frequency stimulation of one pathway (S1) begins (at one-fourth the S0 frequency, Loop = 3). Un-checking the check box of the middle loop (down arrow) stops the lower frequency S1 stimulation. This code is an example of how checking and un-checking a continuous Loop can be used to start and stop different stimulations, and therefore change the protocol flow of execution. Fig. 8b shows a complex LTP/LTD induction protocol. An LTP inducing tetanus was delivered to one pathway (S0) immediately followed by an LTD induction by low frequency stimulation in the other pathway (S1) while the S0 pathway was returned to slow repeat stimulation. The induction stimulation begins when the check box on the Run line is checked. This produces a 1 s, 100 Hz LTP induction train in the S0 pathway by T0sweep, followed by rapid 1 Hz LTD stimulation in the second pathway (S1) by P1sweep, but with continuing S0 low frequency (1/30 s) stimulation from the T1sweep. The P0sweep contains one S0 pulse, the P1sweep contains one S1 pulse, the T0sweep contains a 2 s S0 train, and the T1sweep contains both a single S0 pulse and a single S1 pulse. Note that this LTP/LTD induction protocol cannot merely be produced by evoking LTP stimulation and then LTD stimulation because the slow repeat S0 pulse output continues during the more rapid S1 LTD stimulation output. Fig. 8c shows an example of LTP/LTD stimulation with steady depolarization. If the sweeps have a steady depolarization throughout the sweep, and the Sweep Duration is equal to the Sweep Period (in other words, there is no time between sweeps), then steady intracellular voltage changes can accompany the extracellular stimulation. The LTP/LTD stimulation here shows such steady depolarization when repetitive 0.5 Hz T0sweeps depolarized by 70 mV to produce LTP, and was immediately followed by repetitive 1 Hz T1sweeps depolarized by 40 mV to
354
W.W. Anderson, G.L. Collingridge / Journal of Neuroscience Methods 162 (2007) 346–356
Fig. 8. More LTP/LTD script protocols. (a) Lower stimulation frequency of one pathway (S1) vs. another pathway (S0). The lower frequency S1 stimulation (at one-fourth the S0 frequency, Loop = 3) was started by checking the inner continuous loop (“Loop [99999]”) at the up arrow and stopped by un-checking the Loop at the down arrow. (b) LTP train stimulation in the S0 pathway followed by LTD stimulation in the S1 pathway accompanied by continued slow repeat stimulation of the S0 pathway. Stimulation was started by checking the check box on the Run line (up arrow), and then un-checking it once the LTP/LTD stimulation had started (down arrow). (c) LTP/LTD stimulation with associated steady depolarization. LTP/LTD stimulation was started by clicking the Run check box (up arrow) and then un-checking it when LTP/LTD stimulation had started (down arrow). S0 pulse stimulation is recorded in AD0 and intracellular stimulation voltage from AnalogOut0 is recorded in AD1. P0, T0 and T1 sweeps here have S0 single pulse stimulation, T0sweep has steady 70 mV depolarization, and T1sweep has steady 40 mV depolarization.
produce LTD. This stimulation was started by checking on the check box on the Run line (up arrow), and then un-checking the Run line check box once the LTP/LTD stimulation has started (down arrow).
4. Discussion A versatile LTP program needs to perform several functions including special LTP analyses such as population-spike
W.W. Anderson, G.L. Collingridge / Journal of Neuroscience Methods 162 (2007) 346–356
amplitude, on-line analyses, alternating dual extracellular pathway stimulation, LTP induction by single train, theta burst and primed burst stimulation, and LTD induction by low frequency stimulation. These capabilities are now performed by many electrophysiological programs including pClamp, PatchMaster, AxoGraph, Signal, Nclamp, NAC Gather, WinWCP, and WinLTP’s DOS predecessor, the LTP Program. The present program, WinLTP, has several additional capabilities that makes it particularly useful for the study of LTP and LTD. 4.1. Advanced capabilities of WinLTP First, most electrophysiological data acquisition programs use cursors or stimulation epochs to determine the regions over which synaptic potentials will be analyzed. Therefore, only one or a few synaptic responses are usually analyzed in an acquisition sweep. In contrast, WinLTP, and its predecessor the LTP Program, can analyze up to 1000 S0- and S1-evoked synaptic potentials in a sweep, and in both AD channels. Furthermore, we have extended the analysis of synaptic potentials to those evoked by entire trains of stimulation. The synaptic responses evoked by train stimulation can either be analyzed as many individual synaptic responses, or the train analyzed as a whole, or each individual synaptic response can be measured relative to the baseline of the first pulse. When analyzing the synaptic responses to high frequency trains, we also found that contamination by stimulus artifacts prohibited accurate measurement of peaks and areas, so we added a function to automatically remove the stimulus artifacts. To our knowledge, the only other electrophysiological program with these capabilities is the LTP Program. Second, WinLTP is a multitasking program. Currently it simultaneously runs two independent tasks: the Stimulation/Acquisition Sweeps task, and a Continuous Acquisition recording task. This allows one, for example, to record miniature synaptic potentials or other spontaneously occurring events while at the same time to periodically analyzing and plotting on-line series resistance and cell input resistance. However, it is important to stress that because the tasks are independent, any stimulation/acquisition sweep protocol can be run with the Continuous Acquisition task. Furthermore, while the Continuous Acquisition task is running, the protocol flow, the sweep stimulation values, and the detection values in the Stimulation/Acquisition Sweeps task can still be changed. In other electrophysiological programs, such as pClamp, CED Spike2/Signal, WinEDR/WinWCP, NClamp and AxoGraph, the stimulation/acquisition sweep and continuous acquisition functions can be run separately, but not at the same time. All the standard electrophysiological software is multitasking in the sense that they can generate stimuli while simultaneously acquiring data during a sweep. WinEDR can even generate repetitive pulse or train stimulation during continuous acquisition. However, WinLTP is fundamentally different in that the two tasks can be completely independent of each other (in fact they are run by completely separate threads of execution). It is almost as if two separate programs are simultaneously running but are sharing one data acquisition board.
355
A preliminary non-released version of WinLTP also runs a third task, Capturing Spontaneous Events, which detects and plots spontaneously occurring events such spontaneous epileptiform bursts. However, because this is not fully implemented, it is not available in the current distributed version (0.94). Third, WinLTP has Dynamic Protocol Scripting which enables very complex protocols consisting of Sweeps, Delays, Run lines, and crucially, Loops with counters and loop within a loop constructs to be run. Dynamic Protocol Scripting was designed to be an easy to use, more general method of producing complicated protocols. Furthermore, because WinLTP scripting uses a double linked list database, the protocol flow of execution can be easily changed at runtime, as can the numbers of loops, Sweep and Delay Periods, and all sweep stimulations. To our knowledge, only PatchMaster has a similar method of protocol scripting that determines the protocol flow of execution and which includes loops with counters, and loop within a loop constructs, and can be substantially changed at runtime. A second type of protocol generation, using interpreted or compiled static text script, is performed by Signal. While these protocols can contain loop counters and loop within a loop constructs, they use static text script, and therefore the protocol flow of execution cannot be as easily changed at run time. A third type of protocol generation, using sweep or protocol linking which enables one sweep to link to the next, is performed by pClamp using Sequencing Keys, Signal using Protocol Dialog Boxes, AxoGraph, WinWCP and NAC Gather. These programs can form loops of sweeps, but because they have no loop counter (beyond repeating a single sweep a given number of times), they therefore have no automatic way of exiting the loop (except by user input). They therefore have no loop within a loop construct, and do not have a general method of producing complicated protocols. 4.2. Wider uses for WinLTP in studying synaptic physiology Although WinLTP was particularly developed for studying LTP, its flexible protocol generation, sweep stimulation, on-line analyses and multitasking continuous acquisition make it useful for studying a wide range of synaptic phenomena including ischemia and epilepsy. WinLTP, and its predecessor the LTP Program, are being used in studies of ischemia (Coelho et al., 2006). WinLTP also has several capabilities useful for studying epilepsy. WinLTP can train evoke and record electrographic seizures in 100 s long sweeps. WinLTP cannot only induce electrographic seizures with normal repetitive trains (as in Lothman and Williamson, 1993), it can also apply more complex induction stimulation protocols such as placing single trigger pulses between the induction trains to assess the excitability of epileptiform bursts (as in Anderson et al., 1987). And finally, WinLTP can measure the duration and coastline of stimulus-evoked epileptiform bursts and electrographic seizures, measurements that are particularly useful in epilepsy studies. 4.3. Future additions to WinLTP The multitasking and Dynamic Protocol Scripting coupled with the sweep stimulation/acquisition and special LTP analysis
356
W.W. Anderson, G.L. Collingridge / Journal of Neuroscience Methods 162 (2007) 346–356
capabilities provide a solid foundation for basic and advanced data acquisition techniques with which to study LTP and related synaptic events. However, several capabilities must be added to complete WinLTP’s advanced experimental capabilities. Protocol scripting should include if statements to test protocol flow depending on whether amplitude/slope baseline stability has been achieved. Improved sweep stimulation will include five extracellular stimulation outputs (S0–S4), extracellular stimulation with many sequential trains to directly perform primed burst stimulation, intracellular stimulation with trains and ramps, and increment/decrement of stimulation values such as varying intracellular analog voltage and interpulse intervals. The number of Pulse and Train Sweeps will also be increased to five (P0–P4 and T0–T4). Better analysis of patch electrode series resistance (Rs) will require exponential curve fitting of the capacitative current transient. Most innovative will be implementation of the capture and analysis of Spontaneous Events task to study events such as spontaneous epileptiform bursts and electrographic seizures. Acknowledgements We would like to thank Steve Fitzjohn for reading the manuscript, Neil Bannister for discussions, and Zuner Bortolotto, Stephane Peineau and Tom Sanderson for data in the figures. Supported by the MRC. References Anderson WW, Collingridge GL. The LTP Program: a data acquisition program for on-line analysis of long-term potentiation and other synaptic events. J Neurosci Methods 2001;108:71–83. Anderson WW, Collingridge GL. WinLTP: a Windows data acquisition program for on-line analysis of long-term potentiation and other synaptic events,
Program No. 569.19. 2005, Abstract Viewer/Itinerary Planner. Washington, DC: Society for Neuroscience; 2005. Online. Anderson WW, Swartzwelder HS, Wilson WA. The NMDA receptor antagonist 2-amino-5-phosphonovalerate blocks stimulus train-induced epileptogenesis but not epileptiform bursting in the rat hippocampal slice. J Neurophysiol 1987;57:1–21. Barrionuevo G, Schottler F, Lynch G. The effects of repetitive slow-frequency stimulation on control and ‘potentiated’ synaptic responses in the hippocampus. Life Sci 1980;27:2385–91. Bliss TVP, Lomo T. Long-lasting potentiation of synaptic transmission in the dentate area of the anaesthetized rabbit following stimulation of the perforant path. J Physiol 1973;232:331–56. Bortolotto ZA, Anderson WW, Isaac JTR, Collingridge GL. Synaptic plasticity in the hippocampal slice preparation. In: Crawley JN, Gerfen CR, Rogawski MA, Sibley DR, Skolnick P, Wray S, editors. Current protocols in neuroscience. New York: John Wiley & Sons; 2001. p. 6.13.1–23. Colquhoun D, Sigworth FJ. Fitting and statistical analysis of single channel records. In: Sakmann B, Neher E, editors. Single Channel Recording. London: Plenum Press; 1983. pp. 191–263. Coelho JE, Rebola N, Fragata I, Ribeiro JA, De Mendonca A, Cunha RA. Hypoxia-induced desensitization and internalization of adenosine A(1) receptors in the rat hippocampus. Neuroscience 2006;138:1195– 203. Frey U, Krug M, Reymann KG, Matthies H. Anisomycin, an inhibitor of proteinsynthesis, blocks late phases of LTP phenomena in the hippocampal CA1 region in vitro. Brain Res 1988;452:57–65. Larson J, Wong D, Lynch G. Patterned stimulation at the theta frequency is optimal for the induction of hippocampal long-term potentiation. Brain Res 1986;368:347–50. Lothman EW, Williamson JM. Rapid kindling with recurrent hippocampal seizures. Epilepsy Res 1993;14:209–20. Ogden D, Stanfield P. Patch clamp techniques for single channel and whole-cell recording. In: Ogden D, editor. Microelectrode techniques, the plymouth workshop handbook. 2nd ed. Cambridge: The Company of Biologists Ltd.; 1994. p. 53–78. Rose GM, Dunwiddie TV. Induction of hippocampal long-term potentiation using physiologically patterned stimulation. Neurosci Lett 1986;69:244– 8.