Ephus runs within Matlab (version 2007b and higher; Mathworks, Natick, MA, USA) as a series of scripts, classes, and custom toolboxes. No commercial Matlab toolboxes are required. Data acquisition is through NiMex, a custom toolbox which is included in Ephus. GLib, an open-source, cross-platform utility library, is automatically installed.

The Ephus computer must accommodate the data acquisition boards (via PCI/PCIe slots or USB connectors, as appropriate) and, where needed, a video camera (usually an IEEE 1394 interface connection).

National Instruments (NI; Austin, TX, USA) data acquisition boards are supported (AO, E, and M series). NI driver version NI-DAQmx 8.5.05 or higher is recommended. For electrophysiology and mapping experiments boards such as the following are often used: PCI/USB-6229 and -6259; PCI-6040E; PCI-6713. The number and type depends on the number of channels required. Note that Ephus, through NiMex, is able to support precision-timed DIO (“correlated DIO” in NI terminology). This greatly expands the number of available data I/O channels available.

Axon Instruments amplifiers (Multiclamp 700A and 700B, Axopatch 200B; Molecular Devices, Sunnyvale, CA, USA) are supported. Amplifiers are implemented as object-oriented classes, allowing users to add support for their own amplifier hardware.

Most QImaging (Surry, British Columbia, Canada) CCD cameras are supported (those in the Rolera series are not). Hardware triggering of video acquisition by Ephus is possible and requires a QImaging trigger board.

These items include scanning mirrors, a photodiode and a current preamplifier to read out laser power, equipment to control the duration and power of the laser beam such as a shutter and/or Pockels cell (electro-optical modulator). Alternatively, some laser models allow direct analog power modulation.

Ephus allows easy configuration and display of individual experimental parameters via controls on the graphical user interface (window) for the various programs. However, Ephus also allows rapid (<1 s) “en bloc” switching between experiments as a whole, through the use of settings files and programmable shortcuts, termed “hotswitches”. Each program stores its configuration (i.e., data acquisition parameters, window location) to a settings file on disk. A hotswitch captures the settings for a collection of appropriately configured programs and assigns a named shortcut. Once a desired experimental configuration has been saved as a hotswitch, one can rapidly return to this set-up by pressing the associated hotswitch button. For example, a user might have a hotswitch set-up for a series of current injections to examine threshold and firing response in a given neuron, but then use a second hotswitch to transition to a fixed voltage-clamp protocol for mapping synaptic inputs in LSPS. The hotswitch program (Figure 1A) provides one-click access to these shortcuts, and also provides the means for creating, modifying and deleting hotswitches.

Individual experimental set-ups can be joined into a logical group, which is saved as a “configuration set”. Each user typically maintains a small number (e.g., 1–5) of active configuration sets, each with multiple (e.g., 10) hotswitches. The “File” menu provides commands to save or restore the settings of a particular program, or of all running programs (Figure 1A).

Various modules expose events indicating particular points in program execution; these events can trigger the execution of custom user functions. Typically such functions implement custom data analysis following data acquisition. However, user functions can also change data acquisition parameters and therefore support on-line customization beyond what is available via the graphical user interface.

The xsg (“experiment saving GUI”) controls the saving of data files. All data acquired are bound together in a logical manner and all acquisition parameters are saved in a header. Files are saved as double-precision binary Matlab MAT-files, with the extension *. xsg. A single file on disk contains data and header information from all programs. Files are organized in folders and named in a systematic way that can be customized using the xsg controls. The xsg program is shown, during a slice electrophysiology experiment, in Figure 1B. The experiment number is incremented by the experimenter for each cell recorded, while the acquisition number increments automatically when new data is acquired, for example by the ephys program. A manually incremented “epoch” counter is available, allowing users to define different epochs of data acquisition (typically used in off-line analysis routines).

The current version of Ephus along with instructions for downloading and installation is available at the Ephus web site. It is distributed as open-source software under the “HHMI/Janelia Farm Open-Source License”.

When Ephus is launched, the user is prompted to supply an initialization file. This file establishes the correspondence between hardware channels and data, triggering, and other rig-specific settings – i.e., settings that do not change while running Ephus. A documented model initialization file is provided with the Ephus distribution, which the user can readily modify to suit their rig and experimental requirements.

To start Ephus, the command “ephus” is entered at the Matlab command line. During the first launch, certain programs require minor additional configuration (e.g., specifying which directory contains pulse files).

The flow of events in a generic Ephus experiment is depicted as a flowchart (Figure 1D).

Ephus includes a built-in multi-channel software oscilloscope, the scopeGui (Figure 2). This is convenient for amplifier control and on-line data display. Typically it is used to establish and monitor electrophysiological recordings. The scopeGui applies periodic test pulses and displays basic whole-cell recording parameters. Data are displayed, but not saved.

In the ephys program (Figure 3A), after selecting amplifier channels and assigning pulse patterns, a single trial is acquired by pushing the “Start” button. To repeat the same acquisition multiple times at a particular interval, ephys is placed in “External” trigger mode, the loopGui is set to the desired number of iterations and its “Start” button is then pushed to initiate and control the acquisition (Figure 3B).

It is often desirable to execute a series of pulses differing in one or more parameters. For example, a common paradigm in electrophysiology experiments is to present a family of step stimuli of differing amplitudes. This functionality is provided by the pulseJacker (Figures 3C,D), which allows cycle-by-cycle individual hardware channel specification (which channels will be active, and which pulses they will use). This is a particularly powerful feature in Ephus.

The Ephus layout for a typical electrophysiology recording session is shown in Figure 3.

Many electrophysiology experiments involve a microscope and video imaging. Ephus includes the qcam program for both live display of a video image acquired using a digital video camera (see above), as well as saving single and multiple frames (i.e., movies) for off-line analysis. The qcam controls allow adjustments of exposure, temporal averaging, spatial binning, look-up tables, and region-of-interest zooming (Figure 4A). For saved video frames, the file-naming scheme can be specified manually, or delegated to the xsg program. The qcam program is commonly used in preview mode when establishing giga-ohm seals for patch-clamp whole-cell recordings (Figures 4A,B).

The photodiode program is used to calibrate the light sensitivity of a photodiode that samples the beam power. The imagingSys program is used to calibrate the voltage signals controlling mirror galvanometers. Calibration of a Pockels cell is performed at startup, and recalibration can be done at any time from the mapper.

Pulses are made using the pulseEditor program. In addition, a standard set of electrophysiology-related pulses is available at the Ephus web site. The pulseEditor provides functionality for crafting step pulses, with options for specifying all standard parameters (number of steps, delay, duration, amplitude, etc.) and for simple combinations of pulses. Figure 6 shows the pulseEditor controls (Figure 6A) used to define a train of five pulses and to view the resulting waveform (Figure 6B) – this example makes use of the “additive” option to combine a train of five small negative pulses with a single, larger positive pulse. However, it is also possible to generate more elaborate custom pulses – indeed, any arbitrary waveform – by using the signalObject class included with Ephus. Figure 6C provides an example of such a custom pulse: this “chirp” signal is a sinusoid whose frequency increases linearly with time; it is used to study frequency-specific sub-threshold responses in whole-cell recordings.

Map pattern files specify the sequence of stimulation sites in a grid and are available on the Ephus web site. In addition, the mapper accepts custom map patterns.

A particular collection and arrangement of Ephus modules can be saved as a named configuration. Multiple configurations can be saved. In addition, hotswitches provide a way to change program settings rapidly and conveniently.

Every Ephus program has a menu bar providing at least two standard command items: One menu (“File”) provides access to configuration loading and saving. A second menu (“Programs”) provides access to all programs (Figure 1).

A subset of programs has these buttons. Pushing “Start” immediately triggers the acquisition. Pushing “External” puts the program in external-trigger mode. Programs in external mode can be triggered by pushing the “Start” button of another program; for example, in this way the loopGui can be used to launch ephys, stimulator, and acquirer together, ensuring a common start time. Alternatively, a trigger pulse generated by software or hardware outside of Ephus can also be used in external mode.

In addition to setting the number of iterations and starting loops, the loopGui also provides an option for seamless (DAC mode) or intermittent (CPU mode) looping. In DAC mode a continuous record of the iteration sequence is saved to a single data file, while in CPU mode separate files contain a record for each iteration. The DAC mode provides real-time reliability between loop iterations, at the expense of logging/sending data at all times. Meanwhile, the CPU mode allows for closed-loop control and sparser data-logging, at the expense of potential inter-iteration jitter (typically on the order of milliseconds to tens of milliseconds) and longer inter-iteration intervals.

The overall duration of stimulus pulses is not set in the pulseEditor; instead, the “Trace Length” parameter in the programs that use these pulses (e.g., ephys, stimulator) individually sets the duration of the stimulus pulses. Similarly, this parameter sets the duration of data acquisition in the acquirer.

The “Update Rate” and “Display Width” parameters in the data-acquiring programs (ephys and acquirer) allow users to modify the way data are displayed (there is no effect on the way data are stored). This is particularly useful when acquiring traces that are many seconds or more in duration.

The use of configurations and hotswitches was described above. These allow users to switch rapidly between complex experimental configurations while viewing only those particular program windows necessary to the task at hand.

While the pulseEditor provides controls for defining trains of square pulses, arbitrary stimulus waveforms (see Figure 6C) can be defined using the lower-level signalObject Matlab object included with Ephus. To use a custom stimulus the user first defines the waveform (e.g., as a linear combination of trigonometric functions, as an arbitrary analytic equation in time, a named probability distribution, or from arbitrary array data), saves the resulting signalObject as a pulse file, and selects it for delivery on an amplifier or other hardware DIO or DA channel as specified in stimulator or ephys.

Optical stimulation protocols are not limited to the capabilities offered by the mapper. For example, custom routines can be used on-line to rapidly generate and deliver precise spatial and temporal stimuli to investigate spatio-temporal integration in synaptic circuits.

Ephus provides many hooks in the event flow and elsewhere that allow “user functions” to act on events and data. A number of user functions are built-in. Templates are provided to facilitate development of custom user functions.

There is growing interest in combining two-photon laser scanning microscopy and electrophysiological recording, an approach that has been used extensively for experiments such as high-resolution imaging of Ca²⁺ dynamics in dendritic spines. Ephus and ScanImage can be used simultaneously for this purpose (O'Connor et al., 2008). For most applications, this is as simple as linking the trigger signals between the two software packages, which is often all that is needed to interface Ephus with external instrumentation. More advanced experiments may employ custom user functions to simplify operating both software suites quickly. Telegraphed data, in either direction, may be used to provide timestamps across programs for use in analysis. Depending on hardware availability, physical space, and other concerns, Ephus and ScanImage may be run on the same computer (using one or separate instances of Matlab) or on separate machines.