Analysis of Postural Stability After Concussion Using Empirical Mode Decomposition: A Pilot Study

Maintaining upright stance is a complex process, it requires appropriate functioning of a postural control system which consists of inputs from somatosensory, vestibular, musculoskeletal, and proprioceptive systems as well as from several brain regions [1–4]. A concussion is defined as a brain injury caused due to unexpected acceleration/deceleration of the head causing temporary alteration of brain function and it is a prevalent source of injury to football athletes [1]. With the altered function of the brain, the ability to maintain postural equilibrium becomes challenging due to the inability of individuals to respond promptly to stressors, thus, making maintenance of postural equilibrium rather difficult for individuals with a concussion. Effects of concussion on postural ability are shown to last up to three days post injury [5]. Postural stability test, therefore, can be performed to make a valid return to play (RTP) decision, pre-mature RTP is shown to have been catastrophic due to its potential to permanently impair previously affected region/functioning [1,5]. Postural sway data (center of pressure, COP) is traditionally analyzed to study the postural control. Therefore, COP can provide critical information regarding individual’s ability to maintain upright stance post injury. A more sensitive concussion assessment tool based on electroencephalogram (EEG) is used to accurately track effects of concussion [6]. However, sophisticated electrode placement requirement inhibits its immediate applicability. In current preliminary research, we attempt to differentiate athletes with a history of concussion (experimental) from healthy (control) using postural data. In order to do so, a concept of empirical mode decomposition (EMD) was adopted. EMD has shown evidence in the literature to infer vital information pertaining to the complex underlying physiological phenomenon [4, 7–8]. In the current research, the resultant COP (COPr) was decomposed into its finite set of band-limited signals termed as intrinsic mode functions (IMFs) [8], a set of linear and nonlinear features were extracted from COPr and its IMfs. Lastly, a test of significance was conducted to infer the potential of postural data for differentiating concussed from healthy athletes.