Activity of tongue muscles during respiration: it takes a village?

Alan J. Sokoloff

Almost twenty years have passed since the muscular hydrostat theory of tongue function was set forth (7). Yet few studies have explored one of its central predictions: that “extrinsic” tongue muscles (muscles with origin outside the tongue body) and “intrinsic” tongue muscles (muscles with both origin and insertion in the tongue body) must be coactive in most tongue movements. An exclusive focus on extrinsic tongue muscles is especially notable in studies of tongue movement during respiration for which a single tongue “protrusor” muscle, the genioglossus, commonly serves as the indicator of respiratory drive to the entire tongue. That the situation is substantially more complex is evident from the accompanying study of Bailey and Fregosi (2). The authors demonstrate, in an in vivo rat model, that at least one intrinsic tongue muscle, the superior longitudinal, can be coactive with the extrinsic tongue muscle hyoglossus during respiration. This study thus serves to caution against approaches that reduce tongue movement to functionally discrete extrinsic and intrinsic muscle components, as well as approaches that investigate the neuromuscular control of the tongue in respiration by isolated study of the genioglossus muscle.

The mammal tongue is undeniably complex. Seven or eight tongue muscles, each with extensive terminations in the tongue body, are variously described. In the rat, over 3,000 motor units comprise the muscles of one side of the tongue alone. Tongue muscles are typically categorized according to morphology (i.e., extrinsic versus intrinsic) or to presumed function (i.e., protrusor versus retrusor), and these categorizations form the basis for virtually all studies of the tongue motor system, from cellular to systems levels of investigation. Yet morphologic and kinematic evidence suggests that neither categorization scheme is functionally based or even heuristic. Extrinsic and intrinsic muscles share overlapping courses in the tongue body (1). Thus, with the exception of the extralingual portion of extrinsic muscles, the mechanical effects of many extrinsic and intrinsic muscle fibers are likely similar, not disparate. The same may be said for some protrusor and retrusor muscles (3). Additionally, movements of the mammal tongue are not restricted to the protrusion-retrusion axis but involve complex changes of tongue shape in three dimensions. These changes include the simultaneous lengthening and shortening of different tongue regions (6), a behavior clearly incompatible with a simple protrusion-retrusion model of tongue function.

A conceptual basis for approaching the morphological and kinematic complexity of the mammal tongue is provided by the muscular hydrostat model of tongue function. In this model, the tongue is defined as a constant-volume structure, and tongue movement is thought to be produced by change in tongue body shape (7). Tongue protrusion, for example, is produced by a decrease in the cross-sectional area of the tongue body. The muscular hydrostat model is at its core democratic. The activation of any motor unit with tongue body presence can influence tongue shape and hence tongue movement. Because muscle fiber orientations vary with tongue shape, the mechanical effect of a motor unit (or entire muscle) is dependent on the integrated activity of all other tongue motor units. Multiple muscles work in concert to provide both the structural support and the active force for tongue movement.

These considerations lead to predictions of tongue muscle activity that are contrary to predictions based on traditional schemes of muscle categorization. According to the muscular hydrostat model, protrusor and retrusor coactivation, as well as extrinsic and intrinsic coactivation, should occur in most tongue movements (8). Studies of the extralingual portions of extrinsic tongue muscles, accessible to in vivo electromyographic (EMG) study, have indeed demonstrated that protrusor and retrusor muscles are coactive in respiration (4). But, for most tongue movements, the business end is the tongue body, and here muscle interdigitation limits EMG assessment of individual muscle activity. In the accompanying article, Bailey and Fregosi (2) make clever use of rat tongue anatomy to functionally isolate a single intrinsic tongue muscle, the superior longitudinal, and compare its EMG activity with that of an extrinsic tongue muscle, the hyoglossus, during respiration. Their findings are striking. In the vagus-intact preparation, the hyoglossus is always active, and the superior longitudinal is sometimes active, during normal respiration. At high levels of hypercapnia the hyoglossus and superior longitudinal are always coactive. In the vagotomized preparation, hyoglossus-superior longitudinal coactivation occurs under both normal and hypercapnic conditions. Extrinsic-intrinsic coactivation is thus present during respiration in the rat, and this coactivation is facultative, not obligate. The study of Bailey and Fregosi, in confirming a primary prediction of the muscular hydrostat model, suggests that this model applies to tongue movements in normal breathing and breathing-related disorders such as sleep apnea.

Multiple tongue muscles are thus coactive in many tongue movements, including movements during respiration (2, 4), but these patterns of coactivation are not explained by traditional muscle categorizations. What strategies of neuromuscular activation then explain observed muscle activity patterns? I suggest that the answer to this question requires the in vivo investigation of individual motor units sampled from multiple tongue muscles and will not be obtained from muscle-level investigations. Heterogeneous behavior is expressed by tongue motor units. In the human genioglossus muscle, for example, at least two populations of motor units are present, one population selected into activity during inspiration and expiration and another population selected into activity only during inspiration (10). The neural basis for the selective activation of tongue motor units is not known. Tongue motor units vary with respect to physiological properties (5, 9) and are likely localized to specific regions of tongue muscles (9). It is thus possible that heterogeneous activation of motor units represents a strategy directed within individual tongue muscles to select motor units that share similar properties (e.g., fatigability, speed of contraction, location). But it is equally plausible that selection strategies are directed across muscles to activate motor units from multiple muscles that in concert produce a specific tongue movement (e.g., tongue protrusion). In this scheme, motor units, and not muscles, are the output elements of the tongue motor system. I suggest that cross-muscle strategies are crucial for controlling tongue shape, position, and stiffness so that important behaviors such as breathing, swallowing, feeding, and speech are enacted with maximal mechanical and metabolic efficiency. Moreover, a loss of cross-muscle control of the tongue may initiate or exacerbate conditions such as obstructive sleep apnea; this also underscores the need to focus study on the entire “tongue musculature” as opposed to the activity of only one (e.g., the genioglossus) of the many intrinsic and extrinsic tongue muscles.