Abstract
Breathing is a complex neuromuscular process vital to sustain life. In preclinical animal models, the study of respiratory motor control is primarily accomplished through neurophysiologic recordings and functional measurements of respiratory output. Neurophysiologic recordings that target neural or muscular output via direct nerve recordings or respiratory muscle electromyography (EMG) are commonly collected during anesthetized conditions. While offering tight control of experimental preparations, the use of anesthesia results in respiratory depression, may impact cardiovascular control, eliminates the potential to record volitional nonventilatory behaviors, and can limit translation. Since the diaphragm is a unique muscle which is rhythmically active and difficult to access, placing diaphragm EMGs to collect chronic recordings in awake animals is technically challenging. Here, we describe methods for fabricating and implanting indwelling diaphragm EMG electrodes to enable recordings from awake rodents for longitudinal studies. These electrodes are relatively easy and quick to produce (∼1 h), are affordable, and provide high-quality and reproducible diaphragm signals using a tethered system that allows animals to ad libitum behave. This system is also designed to work in conjunction with whole-body plethysmography to facilitate simultaneous recordings of diaphragm EMG and ventilation. We include detailed instructions and considerations for electrode fabrication and surgical implantation. We also provide a brief discussion on data acquisition, material considerations for implant fabrication, and the physiological implications of the diaphragm EMG signal.
Significance Statement
Investigations of respiratory neuromuscular output have frequently involved the diaphragm muscle, given its role as the main inspiratory muscle in mammals. While anesthetized preparations are fundamental to our understanding of respiratory neuromuscular control, using awake, ad libitum behaving models enhances translatability, particularly when developing and evaluating therapeutic strategies in conditions that result in respiratory pathology and impaired breathing. Here, we describe an affordable and easy-to-produce electromyography implant that enables stable recordings of the diaphragm output in awake, ad libitum behaving rodents over long-term experimental protocols and offer a brief commentary on data acquisition and analysis of these signals.
Introduction
Breathing is a complex, rhythmic motor process involving the coordination of many muscles. The diaphragm is considered the main inspiratory muscle in mammals and also plays a role in other critical nonventilatory behaviors such as cough, sigh, and sneeze (Sieck and Fournier, 1989; Mantilla et al., 2011; Fogarty et al., 2018). The dome-shaped structure of the diaphragm can be divided into costal (ventral), crural (dorsal), and sternal regions (Kocjan et al., 2017). Although all regions of the diaphragm participate in respiratory activity; the costal region is thought to contribute most to normal, quiet breathing (Pickering and Jones, 2002; Nason et al., 2012; Merrell and Kardon, 2013). The diaphragm is controlled by phrenic motor neurons found in the cervical spinal cord, which transmit signals via the phrenic nerve (Ricoy et al., 2019; Fuller et al., 2022). The phrenic nerves insert into the costal diaphragm and provide rhythmic neural drive to facilitate diaphragm contraction (Nason et al., 2012; Fuller et al., 2022).
Preclinical work has investigated diaphragm activity through implanted electromyography (EMG) electrodes with the majority of studies recording diaphragm EMG output under anesthetized conditions (Dow et al., 2006, 2009; Mantilla et al., 2010, 2011, 2013; Seven et al., 2014; Lee and Hsu, 2017; Rana et al., 2017, 2020; Jimenez-Ruiz et al., 2018; Popp et al., 2023; Mickle et al., 2024; Khurram et al., 2024a,b). More recent studies from our laboratories and others have examined diaphragm activity in awake, ad libitum behaving rats (Navarrete-Opazo and Mitchell, 2014; Alvarez-Argote et al., 2016; Bezdudnaya et al., 2018; Jimenez-Ruiz et al., 2018; Rana et al., 2021; Malone et al., 2022; Khurram et al., 2023, 2024a,b; Holmes et al., 2024). While anesthesia offers tight control of experimental preparations, its use can limit the translational nature of the experimental findings. Specifically, anesthesia can suppress central respiratory drive (Mills, 2018; Ballók et al., 2023) and can alter important variables such as cardiovascular and respiratory pattern and reflexes (Hunter and Milsom, 1998). Recording diaphragm EMG output in awake conditions allows for longitudinal studies of the diaphragm muscle during ventilatory behaviors without the confounding effects of anesthesia on neuromotor control. It also allows for nonventilatory behaviors to be recorded including sniffing and sighing. However, given the location of the diaphragm within the body cavity, its rhythmic activity, and the associated risk of pneumothorax, designing and implanting EMGs into the diaphragm muscle present distinct challenges. To address these issues, we designed a low-cost, easy-to-produce EMG method that effectively overcomes these obstacles and achieves successful outcomes (Malone et al., 2022; Holmes et al., 2024). Here, we detail the methods that enable chronic recordings of bilateral diaphragm EMG activity in awake, ad libitum behaving rodents. We follow with a brief discussion on data acquisition, interpretation, and equipment considerations for device fabrication.
Materials and Methods
Animals
All animal procedures were performed in accordance with the Institutional Animal Care and Use Committee at Marquette University (Protocol No. AR-4198) and University of Florida and conformed to the Society for Neuroscience Policies on the Use of Animals in Neuroscience Research. Here we provide instructions to make and implant diaphragm EMGs from our experience with n = 69 male and n = 42 female Sprague Dawley rats. We provide new data [coefficient of variation (CV) for peak EMG, signal-to-noise ratio, and representative images of EMG output] from animals (n = 6 male and n = 7 female) in which EMG signals were recorded for 8 weeks after implant included in our previously published study (Holmes et al., 2024). Additionally, we include new data (signal-to-noise and representative images of EMG) from animals (n = 6 males) in which EMG signals were recorded for 40 weeks after implant that are included in a manuscript currently in press. We also report surgical and postoperative complications from n = 105 of these animals.
Data analysis and statistics
Signal-to-noise ratio was calculated from the integrated channel of diaphragm EMG measurements over a five-breath window of quiet breathing. The signal-to-noise ratio represents the peak amplitude during the inspiratory burst compared with the peak amplitude between each inspiratory burst. CV was determined by taking the average peak amplitude over a 60 s window of quiet breathing across early (4–5 d after implantation) and chronic (8 weeks postimplantation) timepoints. Paired t tests were used to compare the signal-to-noise ratio across 8 and 40 week timelines. The significance level was set to p < 0.05, and all data are presented as standard error of the mean (SEM).
Fabrication of the electrodes
Materials needed for the fabrication of the diaphragm EMG implant are shown in Table 1. A schematic of electrode fabrication and assembly can be found in Figure 1 and is referenced throughout the protocol. In total, it requires ∼1 h to build one implant and to apply the epoxy.
Prepare and clean the individual pieces of the electrode (Fig. 1A).
Pass the top surface of the pedestal over the sandpaper 2–3 times to aide mechanical bonding at later steps.
Spray the gold pins and plastic pedestal with 70% alcohol to remove any residual oils and promote proper solder and epoxy binding. Let the materials dry completely before proceeding.
Cut three 32 cm recording wires and one 16 cm grounding wire.
Four recording electrodes will be implanted into the diaphragm muscle. A single ground wire will be used for this implant.
Each plastic pedestal has six open pinholes. Implants can be constructed using five gold pins (four electrodes, one ground) or six gold pins (four electrodes, one backup electrode, one ground). If utilizing five gold pins, cover the additional, open pinhole with a small piece of tape.
Deinsulate ∼2 mm of the perfluoroalkoxy (PFA) coating from each end of all wires using either a scalpel blade or an open flame (Fig. 1B).
Deinsulating wires with an open flame can leave residue on the metal wire and may affect connectivity and soldering connections at later steps.
Prepare the gold pin and wire connections.
Brush the deinsulated portion of the wire and the edge of the gold pin with a flux pen.
Avoid touching the end of the wire or the pin with ungloved hands.
Solder the wire to the gold pin.
Using a heat-resistant silicone cover on the alligator clip from the helping hands to reduce damage to the pin and increase friction between the pin and the helping hands to improve stability, hold the gold pin in an upright position (Fig. 1C).
Use an additional silicone-covered alligator clip to hold the wire.
Place the deinsulated portion of the wire inside the gold pin (Fig. 1D).
Use enough solder to fill the pin, but not so much that it spills out.
Once completed, cover the top of the pin with solder to create a “dome” to reinforce the wire–pin structure by briefly touching the top of the pin with the solder iron (Fig. 1E).
Repeat until all wires have gold pins at each end.
Use a multimeter to measure the resistance between the gold-plated pins on each recording wire (Fig. 1F).
Expected resistance of recording wire is <80 Ω. Expected resistance of ground wire id <40 Ω.
Note: resistance has a direct relation to the length of the wire (i.e., Higher resistance for longer wires; R = ρ * (l/A) where R is resistance, ρ is the resistivity of the wire material, l is the length of the wire in meters, and A is the cross-sectional area of the wire (in square meters).
Cut each wire in half (∼16 cm for recording wire, ∼8 cm for ground wire).
The length of the wire can be extended to match the size of the animal. We have had success with electrodes up to a length of 25 cm (enough for the wire to reach from the headcap to the diaphragm with additional length to allow for movement and growth of the animal). Adjust the initial length of wire in Step 2 as appropriate if wanting to use longer electrodes.
Prepare the distal portion of the electrode to interface with the muscle. Electrode placement can be either “full-thickness” or “partial-thickness.” Full-thickness placement penetrates through the diaphragm muscle, while partial-thickness placement is embedded within the diaphragm muscle (details below). We have had success with both full- and partial-thickness placement of electrodes. When electrodes are used for partial-thickness implantation, additional assembly steps at the time of surgical implantation are required.
Partial-thickness placement: No additional steps needed at this time.
Full-thickness placement: Tie an anchor knot ∼2 in from the distal portion of the recording wire. Deinsulate the wire from the end to the anchor knot (Fig. 1G). Remove a 31 G needle from a BD insulin syringe. Thread the deinsulated portion of the wire inside the 31 G needle. Crimp the end of the needle where it meets the wire using a flat-blade screwdriver (Fig. 1H).
Be sure that the wire is not exiting the needle at the beveled end.
Assemble the headcap according to the configuration provided in Figure 1I.
Identify the pairs of recording electrodes by tying the suture of varied sizes around each pair (Fig. 1J).
Electrodes which will be placed in the left hemidiaphragm can be denoted by using a small length of size 0 suture.
Electrodes which will be placed in the right hemidiaphragm can be denoted by using a small length of size 3.0 suture.
All recording electrodes are grouped by using a longer length of size 0 suture (note: the longer size 0 suture grouping all recording electrodes is not shown in Fig. 1J for clarity).
Cover the pedestal and gold pins of the implant with clear epoxy for protection using a 1 ml syringe (Fig. 1K). Two options are available to facilitate this step:
If an SLA printer is available, 3-D print the provided headcap holder (Data Repository 1 https://figshare.com/s/c1ae2d737196e34f6b34).
Allow the epoxy to cure for 2–3 min in the syringe before applying over the pedestal. The headcap holder aids in shaping the epoxy which reduces the spillover of the epoxy around the edges of the headcap (Fig. 1K, bottom image).
If no printer is available, secure the pedestal on the mounting holder so that the soldered gold pins are facing up.
Allow the epoxy to cure for ∼10 min in the syringe prior to application. Use a thin wooden dowel to shape the epoxy into a dome structure throughout the remaining cure time (20–30 min; Fig. 1K, top image).
Allow the epoxy to dry for 24 h.
Autoclave the completed implant using settings for dry surgical instruments.
Diaphragm EMG electrode fabrication and assembly. A, Each EMG implant is composed of one plastic pedestal and five or six (if using an extra/backup electrode) gold pins. B, Deinsulate ∼2 mm of wire which will interface with a gold pin. Deinsulated portion of the wire is denoted with the bracket. C, Use the helping hands to hold the gold pin in an upright position. D, With a second helping hand, place the deinsulated end of the wire inside the gold pin. E, Solder the wire and gold pin with enough solder to fill the pin, but not spill out of the pocket. F, Use a multimeter to measure the resistance between gold pins on the same wire. Repeat for all recording and ground wires. Cut each wire in half. G, Tie an anchor knot on each recording wire opposite of the gold pin. Deinsulate the wire from the anchor knot to the distal end. H, Crimp a 31 G insulin needle onto the wire using a flat-blade screwdriver. I, Place each pin into the pedestal according to the configuration. We utilize Pinholes 1 and 2 (CH01) for the left hemidiaphragm, Pinholes 4 and 5 (CH02) for the right hemidiaphragm, Pinhole 3 for the ground, and Pinhole 6 for the extra, backup electrode (if used). It is vital to maintain a consistent configuration among produced implants to ensure correct identification of paired electrodes. J, Once assembled, tie suture around the pairs of electrodes to identify during surgery. Wires from electrodes in Pinholes 1 and 2 (CH01/left) will be identified with a size 0 suture; wires from electrodes in Pinholes 4 and 5 (CH02/right) will be identified with a size 3.0 suture. All recording electrodes will then be tied together with a larger size 0 suture to assist when tunneling the wire under the skin. K, Completed implant is shown with the cured epoxy both with and without using the 3-D–printed headcap holder. If using partial-thickness electrodes, steps found in subpanels G and H will not be performed.
Materials and equipment required to build the electrodes
Implantation of diaphragm electrodes
All procedures described below should be performed under aseptic conditions. Equipment needed for surgical implantation can be found in Table 2.
Anesthetize animal with isoflurane or other preferred anesthetic per individual institution and protocol guidelines.
Once anesthetized, move the animal to the presurgical station and maintain anesthesia in 100% oxygen through a nose cone.
Confirm effective sedation via protocol guidelines.
Prepare the surgical sites.
Shave the dorsal surface of the animal's head and back, beginning caudal from the eyes and extending to the inferior angle of scapulae.
Carefully flip the animal and shave the abdomen from the xiphoid process of the sternum to just rostral of the urethral orifice.
Move the animal back to prone position being careful to maintain the nose in the nose cone.
Clean and disinfect the dorsal surgical area (skull and scapulae) using alternating chlorohexidine and isopropyl alcohol or sterile saline gauze (follow the guidelines of your institution for aseptic practice).
Carefully move the animal to the sterile surgical field in the supine position onto a piece of commercially available cling film (Emmer et al., 2019) or alternative sterile surgical wrap. Be sure to avoid contact with the prepped dorsal surgical area.
Clean and disinfect the abdominal surgical area using alternating chlorohexidine and isopropyl alcohol gauze or sterile saline gauze. Allow the abdomen to dry and then wrap the cling film or surgical wrap around the animal. This method allows surgeons to flip the animal multiple times while maintaining a sterile field throughout the procedure.
For optimal visualization, the animal should be positioned perpendicular to the surgeon (with the tail closest to the surgeon).
Return animal to the prone position.
Fix the implant to the skull via bone screws and dental acrylic.
Make a midline incision over the skull and clear the underlying connective tissue from the skull using scissors and a curette (Fig. 2A).
To further aide mechanical bonding of the dental acrylic, the skull can be lightly scored in a crisscross pattern by using the tip of the scalpel. This step is optional.
Use a 0.5-mm-diameter burr to drill into the skull at specified locations between the bregma and lambda sutures (Fig. 2B).
Place self-tapping bone screws at each location, leaving the uppermost 1.5 mm of screw exposed above the skull (Fig. 2C).
Tunnel the wires from the skull to the abdomen using Adson forceps and fine scissors.
Start at the base of the skull and create a subcutaneous tunnel angled away from spinal midline toward the left scapula using blunt dissection techniques. Once the tunnel extends beyond the scapula, move the animal into a supine position.
Make a midline incision through the skin, but not the abdominal musculature, extending from the xiphoid process to the lower abdomen.
Use the Adson forceps and fine scissors to create a subcutaneous tunnel from the rostral portion of the abdominal incision toward the tunnel extending from the skull.
For best results, the animal should be placed on their right side to connect the two tunnels.
Once complete, advance a Halsted-Mosquito Hemostat through the subcutaneous tunnel from the abdomen to the skull.
Use the Halsted-Mosquito Hemostat to grab onto the implant where the suture is tied around all wires (above the knots). Gently pull the recording electrodes from the base of the skull to the abdominal opening. Once the electrodes are through, move the animal back into the prone position.
Wrap the ground wire around the bone screws at the skull.
The epoxied portion of the implant should be in contact with the exposed skull (Fig. 2D) while avoiding putting too much tension on the ground wire.
Secure the headcap to the skull by using dental acrylic via a 1 cc syringe and 18 G needle.
For best results, dental acrylic should be both underneath and completely around the epoxied pedestal.
Allow the dental acrylic to dry (∼5 min) and clean any acrylic from the surrounding tissue.
Use suture to close the skin caudal to the implant/headcap if needed.
Move the animal back into the supine position for the remaining steps of implantation.
Open the abdominal wall along the linea alba to expose the abdominal cavity and view the diaphragm muscle.
Use Adson forceps to tent the abdominal musculature away from the abdominal viscera prior to cutting.
Open the abdominal cavity from the sternum to the lower abdomen to optimize your viewing window.
For partial-thickness placement of electrodes, wires can enter the abdominal cavity via this midline incision or via a lateral approach. For the lateral approach, pierce the abdominal wall away from the midline with a 16 or 18 G needle. Pass the wires through the needle and into the abdominal cavity. Remove the needle by advancing completely through the abdominal wall while being careful not to contact the abdominal viscera. This lateral approach may serve to decrease irritation at the midline abdominal incision once the animal is awake.
For full-thickness electrode placement, wires are unable to be passed into the abdominal cavity using the lateral approach since the needle is already crimped onto the distal end of each electrode.
Optimize the view of the abdominal cavity by utilizing alligator clips to hold the muscle at each side of the sternum. Be careful not to clip onto the skin as this will make surgical closure difficult at later steps (Fig. 2E).
Lift the clips away from the body to visualize the abdominal portion of the diaphragm.
The liver may obstruct the initial view of the diaphragm muscle and can be carefully moved away from the viewing window with forceps and by cutting the Falciform ligament connecting the liver to the abdominal wall. This step is optional as it may cause more bleeding.
Implant pairs of electrodes into the midcostal region of each hemidiaphragm (Fig. 2F).
Remove the longer length size 0 suture which surrounds all recording electrodes and separate the electrodes for the left hemidiaphragm from those for the right hemidiaphragm. Cover one pair with sterile gauze to prevent the wires from getting mixed up. Remove the suture surrounding the pair that you will implant first.
For full-thickness placement of electrodes, hold the crimped needle at its base using the Halstead-Mosquito Hemostat.
Orient the beveled side of the needle to be nearly parallel to the diaphragm muscle wall. Holding it close to the base will allow for a larger portion of the needle to pass behind the diaphragm muscle.
Insert the needle through the dorsal part of the midcostal diaphragm (ventral to the phrenic artery). Utilize the diaphragm contraction to assist with insertion and advancement of the needle through the muscular wall to avoid excessive force on the muscle. Advance the needle toward the ventral aspect of the animal with the contraction. The needle should span a distance of ∼3–5 mm prior to passing back through the diaphragm muscle to the abdominal cavity (Fig. 2G, dashed lines). Note: do not let go of the needle base until the needle tip has passed back through the diaphragm.
Gently pull the needle until the anchor knot contacts the diaphragm muscle. Tie a second knot on the free end of the wire to secure the electrode against the muscle. Cut and discard the excess wire and needle.
Repeat this process to place a pair of electrodes on each hemidiaphragm in a parallel alignment with ∼4–5 mm between electrodes.
If you included a fifth recording electrode and do not need it for surgery, cut the additional recording electrode above the anchor knot and fold the remaining wire into the abdominal cavity.
For placement of partial-thickness electrodes, detach an insulin needle and bend the tip of the needle to a 45–60° angle.
With the concave aspect of the bent needle facing the surgeon, gently insert the needle into the superficial layer of the diaphragm muscle. Be careful not to completely penetrate the muscular wall.
Advance the needle toward the ventral aspect of the animal for ∼2 mm and push the tip of the needle back through the superficial diaphragm muscle layer into the abdominal cavity. The needle will stay in the diaphragm muscle with both ends exposed to the abdominal cavity.
Using Castroviejo needle holders, grasp the distal tip of the electrode/wire and insert into the beveled tip of the needle. Advance the wire so that it exits at the base of the needle.
Once the wire has extended through the length of the needle, gently hold the needle at its base and remove from the diaphragm. The electrode wire will stay within the muscle in the same path made by the needle.
Pull the electrode toward the opening of the abdominal cavity, and use a scalpel blade to expose the distal 0.5 cm of the coated electrode. Then, tie three square knots using nondissolvable 6″ suture at the end of the electrode.
Pull the electrode in reverse until the suture knots reach the diaphragm muscle.
Either use tissue glue or tie another knot at the point of entry of the wire in order to secure.
Repeat this process to have a pair of electrodes on each hemidiaphragm in a parallel alignment with ∼4–5 mm between electrodes.
Carefully close the abdominal musculature and skin using sterile suture.
We have had success closing the muscle layers of the abdomen using interrupted sutures and closing the skin with subcuticular absorbable suture.
Remove anesthesia and provide postoperative pain medications and antibiotics according to institutional guidelines. Schematic depictions of the implant can be found in Figure 2, G and H.
Surgical placement of the diaphragm EMG implant. A, Create a midline incision over the skull, and clear the underlying connective tissue for a clear view of sutures on skull. B, Schematic depiction of location for screw placement. Use a 0.5-mm-diameter burr to drill into the skull at each location marked by “X”. C, Place self-tapping bone screws at each location, leaving the uppermost portion of the screw exposed. D, Once wires are tunneled from skull to abdomen, place implant epoxy-side down onto the skull between the screws. E, Use dental acrylic to secure the implant in place. Dashed lines indicate wires underneath the dental acrylic. F, Using the helping hands, tent the abdomen to view the diaphragm muscle. Insert electrodes at the black X's and advance behind the diaphragm and then back out. Gently pull until the anchor knot is touching the diaphragm and tie a second anchor knot. G, Repeat until two pairs of electrodes are placed in the diaphragm muscle. H, Schematic of the completed diaphragm EMG implant with subcutaneous wires extending from the secured implant to the abdomen and into the diaphragm muscle. Recording cables are plugged into the implant to record the EMG output. Dashed lines show subcutaneous wires from EMG implant to abdomen. Created in BioRender. Streeter (2024) BioRender.com/o298768.
Equipment and tools needed to implant the diaphragm EMGs
Data acquisition
To record diaphragm EMG signals, the data acquisition system requires a differential amplifier, an analog-to-digital conversion interface, and a computer for the control of the recording software, data visualization, and storage of raw data. The use of a digital signal processing unit can be included in the equipment configuration for real-time pre- and postprocessing of the recorded EMG signal. The data provided here used the following settings on the differential amplifier: gain of 1K and a bandpass filter with a low cutoff of 0.1 kHz and high cutoff of 5–10 kHz. The sampling rate was set to 25 kHz. At a minimum, sampling frequency should be set for at least 2.5 times the expected maximum frequency of the signal to reduce the risk of aliasing (Nyquist, 1928; Shannon, 1949; Jerri, 1977; Merletti and Muceli, 2019). Diaphragm EMG signals from the implant are carried to a commutator to enable animal movement while preventing cables from becoming tangled (6 Channel System Commutator, Part SL6C, Protech International) through a six-pin cable (6 Channel Cable with Spring, 363-363, Protech International). A second six-pin cable (6 Channel Cable with no Spring, 363-491/6, Protech International) then carries the signal to an adapter (BNC female to double biding post; A-M Systems 726034) to convert to a BNC cable with a five-pin cable connector (A-M Systems 521500) which plugs into the amplifier. EMG recordings should be stored in its raw format to avoid data loss or corruption. Readers are referred to McManus et al. and Jonkman et al. for detailed information on terminology, setup, and data acquisition of EMG signals (McManus et al., 2021; Jonkman et al., 2024). This system can be utilized in conjunction with both head-out and whole-body plethysmography (WBP rat chamber tower and WBP tower comm lid, Data Science International) to allow simultaneous recordings of diaphragm EMG and ventilatory parameters.
Results
Financial and clinical outcomes of diaphragm EMG implant
The diaphragm EMG implant described here costs ∼$70 to produce and takes ∼1 h to assemble. As included in Table 1, fabricating these electrodes requires standard laboratory equipment. The EMGs can be implanted in roughly 2 h or less, and as shown in Table 3, we calculated the incidence rates of surgical and postoperative complications from n = 105 EMG implant surgeries and provided steps to reduce adverse events associated with each complication. In our hands, adverse complications that interfere with the ability to record signals are relatively infrequent, including a 2% risk for pneumothorax and loss of diaphragm EMG signal on one side in 5% of implanted animals.
Surgical and postoperative complications and incidence rates associated with implanting chronic diaphragm EMGs and suggested steps to reduce each complication
Data processing pipeline for diaphragm EMG signals
The first steps in analysis of the EMG recording are to extract the linear envelop by denoising the raw signal, rectifying, and applying a low-pass filter to the raw diaphragm EMG signal. Traditionally, this has been performed by using a moving average filter, moving root mean square (RMS), or a low-pass filter. To facilitate data analysis of diaphragm EMG signals collected with this implant, we have provided a Python code to extract the linear envelope by each of these methods (Data Repository 2 https://figshare.com/s/c04c07ecf06424e4e56a). NEO (Garcia et al., 2014) was used to load the data efficiently from CED data files. Examples of raw and processed diaphragm EMG signals using this data pipeline are shown in Figure 3. We calculated the CV of the peak EMG amplitude for each processing method and show that the CV for each method (normalized moving average, 14.75%; normalized moving RMS, 7.22%; normalized low-pass filter, 6.93%) aligns with published CV from diaphragm EMG (Mantilla et al., 2011). The instructions for key steps, a link to GitHub, and annotations of the original code used throughout the pipeline for EMG signal integration are provided in the linked repository. Additional programs may be needed for more sophisticated analyses or handling significantly long recordings. It is important to note that each of these approaches is susceptible to noise, movement artifact, and electrocardiogram (ECG) artifact given the location of the diaphragm muscle relative to the heart. Specifically, we see a more prominent ECG signal on the left hemidiaphragm EMG recording given the proximity of the electrode to the cardiac tissue. In our experience, the ECG signal is position dependent and is more prominent when animals are in a crouched/flexed or resting position as compared with locomotive and upright positions.
Examples of processed diaphragm EMG signals. A, B, A representative, preprocessed unilateral diaphragm EMG trace during eupneic breathing and examples of varied data processing approaches for the signal across a compressed (A) and expanded (B) timeline. CV was calculated from the processed data prior to normalization and is noted for each data processing approach for the compressed timeline. RMS, root mean square.
Diaphragm EMG signals during ventilatory and nonventilatory behaviors
Using the provided data pipeline, we show representative raw and integrated EMG recordings from the left and right hemidiaphragm in awake, ad libitum behaving animals (Figs. 4, 5). Specifically, we provide expanded traces to show both ventilatory and nonventilatory behaviors including eupneic breathing, sniffing, and a sigh (Fig. 4). A distinct waveform is observed in the raw diaphragm EMG during each behavior. In addition, we include an example trace of diaphragm EMG activity taken from an animal while in their cage showing transitions between quiet resting breathing, breathing during locomotion, and sniffing (Fig. 5). These data show the ability of our implanted electrodes to capture both ventilatory and nonventilatory behaviors during rest and active locomotion.
Diaphragm EMG signals across respiratory behaviors. A, Representative raw bilateral diaphragm EMG traces for an awake, ad libitum behaving rodent across ventilatory and nonventilatory behaviors using the described implant. B, Integrated bilateral diaphragm EMG signals using the provided data pipeline for an awake, ad libitum behaving rodent across ventilatory and nonventilatory behaviors.
Bilateral diaphragm EMG across respiratory behaviors in an awake, ad libitum behaving rodent. A, Representative diaphragm EMG traces from the left and right hemidiaphragms during a 1 min recording in a ad libitum behaving animal showing various behaviors including (B) quiet breathing, (C) breathing during locomotion, (D) sniffing, and (E) return to quiet breathing.
Stability of diaphragm EMG signals
We routinely record diaphragm EMG signals for several months (Holmes et al., 2024), and signals have remained viable for at least 40 weeks following implantation (Fig. 6). As reported in Holmes et al. (2024), we did not detect differences in impedance after 8,weeks in male or female rats. To further examine the stability of these electrodes, we calculated the signal-to-noise ratio and did not detect differences across 8,weeks (n = 13; p = 0.4551, paired t test) or 40,weeks (n = 6; p = 0.4357, paired t test) postimplant. We also determined the CV for peak EMG amplitude during resting breathing 4–5 d after surgery and show a range of 4.5–14.4% with a mean (±SEM) of 7.04 ± 0.73% (n = 13). Across 8,weeks, intra-animal CV ranged from 3.1 to 7.1% with a mean of 4.803 ± 0.35% (n = 13). These data support the long-term stability of our electrodes to capture diaphragm EMG signals for recordings over several months.
Diaphragm EMG recordings over prolonged time periods. Representative diaphragm EMG recordings from (A) the baseline, (B) 2 weeks postimplant, (C) 4 weeks postimplant, and (D) 8 weeks postimplant. Representative diaphragm EMG recordings from a subset of animals have been collected during (E) the baseline and (F) 40 weeks postimplant. Note: panels A–D were collected with an amplification of 1,000×, and panels E and F were collected with an amplification of 100×.
Simultaneous measurements of plethysmography
Using this approach, animals can be placed in a modified whole-body plethysmography chamber to simultaneously collect multiple respiratory variables. The modified plethysmography chambers use a commutator to facilitate animal movement and prevent cables from becoming tangled. We provide an example of simultaneous recordings of diaphragm EMG output and whole-body plethysmography during normoxia and a maximal respiratory challenge (Fig. 7). As expected, during respiratory challenge, both tidal volume and diaphragm EMG output are increased.
Simultaneous collection of diaphragm EMG and whole-body plethysmography. Representative respiratory flow and right diaphragm EMG traces during (A) normoxia and (B) a maximal hypoxia (10.5%) and hypercapnia (7%) challenge.
Discussion
Diaphragm EMG recordings have been essential in building an understanding of respiratory neuromotor control (Sieck and Fournier, 1990; Goshgarian, 2009; Mantilla et al., 2010; Seven et al., 2014; Del Negro et al., 2018). Recording diaphragm activity in awake, ad libitum behaving animals is crucial to deepen our understanding of basic respiratory function and improve translation of preclinical data. This approach eliminates the confounding effects of anesthesia, allows for nonventilatory but physiologically important behaviors to be captured, facilitates longitudinal studies of diaphragm EMG signals, and enables simultaneous collection of ventilatory data via plethysmography. While telemetry-based EMG systems compatible with the diaphragm muscle are commercially available, these systems do not allow investigators to adjust recording parameters, may have sampling limitations, have limited battery life and unknown rates of success, and carry significant costs. Therefore, we provide detailed instructions on preparing and implanting affordable diaphragm EMGs for chronic recordings in awake animals. The diaphragm EMG implant detailed here costs ∼$70, takes about an hour to assemble, and uses standard laboratory equipment to produce. In addition, the procedures needed to implant these EMGs carry a minimal risk for adverse events and can be implanted in roughly 2 h or less. Although these methods were developed for the diaphragm, our group has also had success in other respiratory muscles such as the geniohyoid (Keilholz et al., 2023), external intercostals, and genioglossus (unpublished observations).
We selected stainless steel for our EMG wires due to its mechanical stability, resistance to corrosion (Bekmurzayeva et al., 2018), and previous success (Dow et al., 2006; Mantilla et al., 2011; Malone et al., 2022; Popp et al., 2023; Holmes et al., 2024; Mickle et al., 2024). Given the propensity of stainless steel electrodes for biofilm formation (Bekmurzayeva et al., 2018), monitoring the signal-to-noise ratio and impedance can be performed as an index of the tissue–electrode interface (Cogan, 2008; Gaire et al., 2018; Mierzejewski et al., 2020; O’Sullivan et al., 2024). In our hands, electrode impedance is stable across days (Malone et al., 2022) and months (Holmes et al., 2024) postimplant. Additionally, we provide new data that suggest our approach provides a high signal-to-noise ratio which does not significantly change over time. Finally, our data show that chronic diaphragm EMGs have a low CV over time. Together these measurements suggest this approach produces stable and reliable diaphragm EMG signals.
Due to the proximity of the diaphragm to the heart, ECG signals can also be present in the diaphragm EMG recordings. While ECG signals can be removed from EMG traces by a variety of methods (Taelman et al., 2007; Zhan et al., 2010; Nitzken et al., 2013; Estrada et al., 2017; Jonkman et al., 2021; Liang et al., 2023; Wang et al., 2023) or through the purchase of commercially available ECG-blankers, this removal may not be necessary given the regularity of ECG burst shape, frequency, and amplitude (Jensen et al., 2017). To date, we have had success handling ECG and other “noise” in our EMG recordings through the use of low-/high-/bandpass filters, notch filters, and comb filters. Additionally, there is substantial published work highlighting methods to mitigate the noise captured through EMG recordings (Lijun, 2005; Boyer et al., 2023; Halidou et al., 2023). The provided data processing pipeline offers a basic entry point for the transition of raw recordings to the linear envelop. While a formal discussion regarding pre- and postprocessing of data is beyond the scope of this manuscript, the reader is referred to recent literature reviews for more in-depth discussions on methodologies (Carvalho et al., 2023; Jonkman et al., 2024).
While not currently used by our group, the use of hydrogel-based adhesives may offer an advantage for implanting the electrodes by reducing tissue damage resulting from insertion of the needle (Zhou et al., 2023). Furthermore, the potential use of tissue adhesives may aid in placement of electrodes without compromising the muscle membrane and offer promise for surgical closure of tissue layers (Yuk et al., 2019). Advancements in materials science present an opportunity to optimize electrode design with improved impedance responses through the use of metal–polymer compositions and increased channel density to allow for a more in-depth study of the motor unit. For example, the Myomatrix electrode by the CAMBER group is a multichannel electrode that allows for high-density recordings of muscle activity via a flexible probe (Chung et al., 2023). Similar systems can be applied to the diaphragm muscle and enable more in-depth investigations into respiratory neuromuscular control in both healthy and pathological conditions.
Diaphragm EMG recordings from awake animals can offer important insights and informative data for understanding respiratory neuromuscular control. It should be noted that the EMG signal does not explain the mechanical aspects of contraction and only reflects the electrical events of a muscle contraction (Roberts and Gabaldón, 2008). Collectively, the data provided here shows the utility of this method to collect multiple outcome measures in each animal over long experimental timelines. This approach is particularly useful to characterize breathing deficits and potential therapeutic treatments during diseases such as spinal cord injury, Pompe disease, Duchenne muscular dystrophy, or other conditions in which breathing may be impaired. For example, we have used this method to simultaneously examine the bilateral diaphragm EMG output and ventilation during normoxia and respiratory challenge following spinal cord injury (Holmes et al., 2024). Additionally, when implanted before spinal cord injury, the implanted wires can be used to verify or validate functional deficits during and immediately after the injury. Diaphragm EMG activity can also be used to classify basal breathing from nonventilatory behaviors such as sniffing through analysis of EMG burst frequency (Moore et al., 2013; Bezdudnaya et al., 2018) or with approaches such as machine learning. Recently, Obaid and colleagues provided a model which incorporates multiple components of the diaphragm EMG burst such as respiratory cycle duration and peak amplitude to automate the identification of eupneic breathing (Khurram et al., 2024a,b). This method would be highly beneficial due to the potential to collect large volumes of data during longitudinal studies.
Common analyses applied to diaphragm EMG recordings include amplitude, time-to-peak, and onset/offset duration of the EMG signal. Amplitude assesses the magnitude of muscle activity during a contraction and can be used to determine the relative change in a recording over time. However, the ability to make comparisons across subjects or across multiple recordings is reliant on data normalization (Law et al., 2011; Mantilla et al., 2011; Jonkman et al., 2024). Several normalization methods are common, including the following: (1) to normalize diaphragm EMG amplitude relative to a sigh or “near maximal” breath (Mantilla et al., 2011; Hernandez-Torres et al., 2017; Rana et al., 2017; Seven et al., 2018), (2) to preintervention (Rana et al., 2021, 2024), or (3), in the case of spinal cord injury, to preinjury amplitude (Navarrete-Opazo et al., 2015; Rana et al., 2021; Sieck et al., 2021; Fogarty et al., 2023). These methods reduce intra-animal variability and improve the quantitative assessment of EMG signals over longitudinal studies. Time-to-peak measures the time duration from burst initiation to peak of the EMG signal and can be useful to assess neural drive (Jonkman et al., 2024). For example, calculating the RMS value of the diaphragm EMG signal 75 ms (RMS75) after the onset of activity can be used to estimate respiratory neural drive (Seven et al., 2014). Finally, onset/offset duration can be used to determine the duration of muscle activation during a contraction or throughout a recording period (Jonkman et al., 2024). Importantly, no clear definition exists for EMG onset/offset given that muscle activation is a multifactorial, integrated, and dynamic process. Thus, identifying the onset/offset is dependent on either automated (Roesthuis et al., 2020; Rodrigues et al., 2021) or manual (Koopman et al., 2018; Epiu et al., 2021) thresholding techniques and varies widely across labs. Examples of EMG data analysis can be found for both anesthetized (Seven et al., 2013, 2014; Gill et al., 2015; Lee and Hsu, 2017; Rana et al., 2017; Khurram et al., 2022) and awake (Bezdudnaya et al., 2018; Jimenez-Ruiz et al., 2018; Rana et al., 2021; Khurram et al., 2023) recordings.
Given the diaphragm muscle's unique characteristics, such as rhythmic contraction and the risk of pneumothorax, careful consideration of electrode materials and surgical procedures can improve the success of implanting diaphragm EMGs. The methods and resources provided here allow the scientific community to produce affordable diaphragm EMG implants that consistently produce reliable and stable signals in awake, behaving animals. The use of diaphragm EMGs during wakefulness offers an avenue for enhanced understanding of respiratory neuromuscular control and improved translation of preclinical studies.
Footnotes
The authors declare no competing financial interests.
This work was supported by funding from National Institute of Health Grant Numbers National Heart, Lung, and Blood Institute R01HL153102 (to E.A.D.), T32HL134621 (to BREATHE Center, Mitchell PI; A.R.M.), and SPARC OT2OD023854 (to Bolser/E.A.D.);
National Heart Lung and Blood Institute, R00HL143207 (to K.A.S.); National Center for Advancing Translational Sciences, UL1TR001436 (to T.C.H.) and 2TL1TR001437 (to T.C.H.); Craig H. Neilsen Foundation Pilot Grant 993667 (to K.A.S.); and the Craig H Neilsen Foundation (pilot grant to E.A.D.).
↵*T.C.H. and J.D.P.-A. are first co-authors.
This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International license, which permits unrestricted use, distribution and reproduction in any medium provided that the original work is properly attributed.
References
In this issue
