The relationship between body posture , gait biomechanics and the use of sensory insoles : a review

Objective: This study is a literature review aimed at synthesizing information regarding the relationship between body posture, gait biomechanics, and the use of sensory insoles, as well as contributing to the investigations on this topic. Methodology: We have collected the research data from the databases Science Direct, MEDLINE/PubMed, Web of Science and Scielo. We used the following descriptors in the search for the articles: body posture, running, injury, plantar pressure, sensors, and sensory insoles; we have also associated these terms with one another in our search. Results: We have selected the articles that contained literature reviews, treatment, or on-site surveys, published up to 2020. This review has identified the existence of several commercially available pressure sensors, with technologies such as capacitive, resistive, piezoelectric, and piezoresistive sensors. This study has also identified several advantages in the use of the insole technology: improvements in balance and speed rates in the anteroposterior region; redistribution of plantar pressure during walking for diabetic patients; alteration of the pressure over time relationship throughout the entire plantar region. Conclusion: The progress obtained by the use of these sensors over the past few years has been motivating researchers to aim for improvements in its performance and practicality, Research, Society and Development, v. 9, n. 9, e263996793, 2020 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v9i9.6793 3 allowing for its use in diagnosing balance disorders, which can be related to body posture and gait biomechanics.

cause or stem from abnormal plantar pressure on the structures of the foot, as well as aid in the development of better strategies for training and treatment (Tas & Cetin, 2019). Therefore, studying plantar pressure is fundamental to understand those mechanisms and injuries and to explore interventions and training that can improve recovery and avoid future injuries.

Sensors and the plantar pressure signal acquisition system
The biofeedback technique was developed approximately 50 years ago and can be used for several medical disorders, including application in physical education activities. For motor learning, the biofeedback methodology is very simple: the professional positions one or more sensors on outer devices across the body to measure specific physiologic processes (Moss et al, 2003). The biologic signals are electronically processed and sent back through auditory or visual feedback. This way, the user can be more aware of the partial or complete evolution of the biologic processes (Pitta et al., 2006;Franklyn-Miller, 2014).  (Orlin & Mcpoil, 2000). There are usually several sensors in multiple areas of these sensory insoles, providing pressure distribution data, given that the pressure distribution pattern is a predictor of gait or body instability (Tao et al., 2012).
The most commonly used pressure sensors measure the plantar pressure distribution across the foot (Ayena et al., 2018). Every loss of balance, slip, stumble, sudden crouching, and ankle twist is associated with surface-specific environmental conditions, such as slippery Development, v. 9, n. 9, e263996793, 2020 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v9i9.6793 9 floors, irregular surfaces or obstacles along the way, which create a plantar pressure exclusively on the foot, generating distribution patterns that can be measured with sensory insoles. Learning algorithms classify the loss of balance events using spatial and temporal resources that reflect data patterns exclusive to plantar pressure (Brassard et al., 2012;Antwi-Afaria et al., 2018).
The progress obtained by the employment of sensors over the past few decades motivates researchers to improve the performance and practicality of sensory recognition in multiple realistic environments (Cornacchia et al., 2017). This is a complex process that follows five essential steps: 1) selecting and implementing the appropriate sensors on the human body or on the environment in order to capture the user's behavior or alterations in the environment where the user is training; 2) collecting the data from the sensors and processing it in accordance to the activity performed by the user; 3) extracting useful resources from the sensor data for classification; 4) training the classification models with the appropriate machine learning algorithms to infer activities; and 5) testing the learning models for performance reports (Lara & Labrador, 2013;Nweke et al., 2018).

Sensory receptors
Professional athletes need their nervous system to quickly recognize and adjust the position of their limbs and joints to practice the sport more efficiently (Ducic et al., 2004).
Afferent signals from the peripheral nervous system send somatosensory information to the central nervous system, where they are processed along with visual and vestibular information to orient the execution of agile and coordinated actions (Peterka, 2002).
The somatosensory system detects the subtle movement of the lower limbs (Horak,