Historically, the integration of physiologic signals coming from a variety of sources has required numerous pieces of equipment. Not only were such setups cumbersome, often taking up multiple racks or cabinets, they exposed the patient to excessive physiological loading while wearing various sensors and apparatuses. In addition, medical engineers and researchers faced the daunting task of finding a way to synchronize the outputs of all these machines with sufficient precision.

Thanks to today’s compact PCs and video recorders, the three modules that make up the full Kay Swallowing Workstation fit easily on a cart that doctors or clinicians can roll to a patient’s bedside.Clinicians working with patients suffering from dysphagia (the medical term used to describe difficulty in swallowing) are now relieved of these problems. Taking advantage of PC technology, data-acquisition boards and the Windows operating environment, Kay Elemetrics Corp developed the Model 7100 Swallowing Workstation. This instrument fits on an easily movable cart yet permits synchronization of several physiological signals along with realtime display as well as videotape recording and digitization during swallowing evaluations. Indeed, the ability to view what’s going on inside a person’s mouth and throat while examining time-synchronized signals from specialized sensors gives physicians, researchers and speech pathologists working in acute care and rehabilitation facilities capabilities until recently unavailable in any form.
Being able to coordinate physiological signals and live video is of great advantage in diagnosis and treatment. In examinations and therapy, clinicians sometimes ask patients to execute complex and abstract tasks using muscle groups that previously operated under automatic response. Execution of these tasks is frequently not observable to the naked eye and, as such, it’s difficult to measure success or improvement. In rehabilitation or physical therapy, muscle retraining is frequently a visible process whereby patients rely on direct feedback to adjustments in motor behavior. Unfortunately, this feedback hasn’t been available in swallowing rehabilitation.
Digitizing sensor signals
The Model 7100 workstation consists of three main modules, all housed in the same rolling cabinet, each of which can function independently or in sync with the others. The first, the Swallowing Signals Lab (SSL), handle digitization of the sensor signals. For that task it employs a Win-30 board from United Electronic Industries, a card that features 16 analog inputs with 12-bit resolution and a peak digitizing rate of 1 MHz.
To measure various parameters, clinicians attach color-coded sensor leads to the workstation and on the patient. Most of the sensors inputs allow sampling rates from 250 Hz to 1 kHz on each channel. The clinician sets the rate as appropriate for the exam.
Supplied standard with the SSL are several sensors vital to studying dysphagia-related problems. First come dual surface electromyography signals for monitoring of the electrical functions in two separate muscle areas. The clinician normally attaches these sensors externally on top of the neck muscles that are critical to swallowing. Being able to view these muscles’ electrical activity during swallowing can assist a patient in learning compensatory techniques to enable him to swallow safely.
Another key parameter concerns respiration, and to measure it the system comes with a small tube (cannula) that sits in the patient’s nostrils. Because the processes of breathing and swallowing share a common pipe (the pharynx), the coordination between breathing and swallowing is finely timed and in most people occurs automatically. However, patients with strokes and other neurological problems can experience problems with this coordination. These signals help patients to monitor and gain volitional control over respiration in relation to swallowing.
Of course, the tongue is key during swallowing, so the workstation must take suitable measurements. For this purpose it supplies a tongue manometer, a multichannel array of air-filled bladders that sit in the patient’s mouth. The five digitized channels monitor tongue strength along with symmetry of applied pressure. Symmetry might be an important variable such as after a stroke if a patient suffers from severe weakness on one side of the tongue that results in poor transfer of food to the pharynx.
Measuring pressure is also important further down the food channel. Thus the SSL supplies a manometer in the form of a solid-state catheter that slips into the patient’s nose. It provides as many as six channels that show the timing and amplitude of pressure events in the pharynx and upper esophageal sphincter during swallowing. These measurements normally take place with concurrent video fluoroscopy, and the workstation allows precise time alignment between the physiological measurements and video data.
The two final inputs to the SSL supply variable sampling rates, and these higher bandwidth signals require an increase in the maximum sampling rate to 8 kHz. One input reads a signal from a stethoscope acoustic microphone placed on the neck to hear the sounds of swallowing and breathing after the swallow. This acoustic signal also serves as an unambiguous marker of the swallow when viewed in conjunction with other types of data. Finally, to handle whatever unusual situations might arise, the workstation comes with two auxiliary channels that allow concurrent viewing of signals from transducers of the researcher’s choice.
The actual combination of signals used during an exam is software selectable. For all the inputs, the workstation provides control of signal gain as well as bandpass filtering to remove noise. The system digitizes these sensor inputs and displays results in real time for immediate examination and visual cueing. The software also allows post-acquisition measurements, storage or printouts that can track progress over time or build a patient database.
Two additional modules
The second module in the system is an FEES (Fiberoptic Endoscopic Evaluation of Swallowing) system for video feedback and recording of activity in the throat and foodpipe. It consists of a 300W light source, flexible endoscope attached to a CCD camera, VGA monitor, SuperVHS video recorder and color printer. In a FEES procedure, the clinician passes a fiberoptic laryngoscope through the nose, positions it close to the larynx and watches key physical mechanisms during the ingestion of drink and food or, in speech therapy, while the patient utters various sounds.
The third module implements a computerized video system for the detailed review and study of swallowing examinations. Using components of the other modules when possible, it consists of a PC specially configured to control the VCR, VGA monitor, customized software and color printer. This module enables a clinician to archive swallowing exams and digitize swallowing images for storage, enhancements, annotation and measurements. The system provides a favorable videofluoroscopic image recording that’s free from substantial distortion during freeze-frame and slow-motion analysis.
Real-time video/signal correlation
A special feature provides the ability to quickly correlate findings in the videofluoroscopic exam with signals from the Signal Lab module with precise, frame-specific accuracy. This capability is particularly useful for clinicians administering manofluorographic studies or researchers wishing to correlate other physiological signals with the video recording.
Indeed, this special capability is one reason that the system’s development team chose the Win30 data-acq board from United Electronic Industries (Watertown, MA www.ueidaq.com), relates software engineer Pedro Davila. When he was writing the initial software, the most advanced operating platform was Windows 3.1. He examined digitizer cards from all the leading suppliers and found that they could all function under Windows. However, Pedro needed results from sensors in real time so he could provide instant feedback to the researchers and also correlate results with the video system from the third module. Unfortunately, most commercial boards use large on-board buffers for temporary storage of acquired samples. This feature, although desirable in some applications, automatically disqualified those products from use in the Kay Swallowing Workstation. In the medical field, it is essential that the acquired data be available for display in real time. The Win30 efficiently moves data from the on-card buffers to the application’s display buffers in host memory. The card’s efficiency in this regard, especially under a Windows environment, was essential in order to provide plenty of spare CPU cycles for the execution of the rest of the application. Kay’s development team wrote their application in C for Windows 3.1 Support leads to successKay credits the success of this project to the support the firm received from UEI. The following example illustrates this support. A critical requirement for the Swallowing Workstation is the realtime display of acquired physiologic data from the Win30 card. More specifically, the system must present the clinician with a display of signals as they’re captured, an aspect especially critical in many medical applications. The original design of the Win30 card and its driver saved acquired data in a 1000-sample FIFO, but the driver transferred data from the FIFO to the application’s memory space only on the Half-Full FIFO condition. In other words, the FIFO would accumulate 512 samples before it would interrupt the PC and transfer the data to the application. This delay was unacceptable for an application such as the Swallowing Workstation that requires realtime data display. Assume, for instance, you’re digitizing a sensor at 250 Hz, a low-frequency signal typical in physiologic systems. With the Half-Full FIFO requirement, some of the physiologic data.