Learn More - Body Composition Models

Learn more about the new options for Styku body composition in Styku version 4.1 and later.

Introducing Styku Phoenix, a new intelligent body composition prediction model. Through partnerships with the United States's National Institute of Health (NIH), our academic partners, and our rock star data scientists, we've completely revamped our prediction models to increase accuracy, and provide more detailed predictions. By aggregating data from populations around the world and well-validated devices, including the gold standard DEXA scanner, we've built robust prediction models to assist you in providing a better experience to your customers. In fact, Styku has developed multiple prediction models for body composition based on your needs and giving you the choice to choose the model that's best for your business.

How do they work?

Styku uses circumferences, areas and volumes of various regions of your body to predict your body fat as measured by well known devices on the market. Since these devices vary considerably, we give you options for choosing two classes of devices. Our advanced model is the most accurate model, as its built using DEXA data. DEXA is considered a medical grade body composition analyzer and is widely considered a gold standard in measuring body fat. But be careful using this method, as it will give you higher body fat values than less accurate and more popular methods today. The other model you can choose is the basic model. The basic model is based on data aggregated from Bioimpedance devices, which typically underestimate body fat% by 5-10%. If your customer, patient, or member is more likely to respond better to a BIA device, then use the basic model.

Advanced Model (Recommended)

This is the most accurate model and is well correlated with DEXA data. By choosing this model, you'll be able to also provide bone mass, lean mass, visceral adipose tissue (VAT), subcutaneous adipose tissue (SAT), Android Fat Mass, and Gynoid Fat Mass predictions to your customers. However, it should be well understood that typically the estimate for body fat are anywhere from 5-10% higher than what you'd expect from a Bioimpedance device or calipers.

Output: Fat Mass, Bone Mass, Lean Mass, Visceral Adipose Tissue, (VAT), Subcutaneous Adipose Tissue (SAT), bone density, Z-score, Android Fat, Gynoid Fat, and A/G ratio.

Basic Model

BIA (Bioelectrical Impedance) devices pass electric current through your body and measures the magnitude by which your body resists it and uses statistical algorithms to correlate it with various body composition metrics. Expect a lower and underestimated body fat with this method If your customers are used to seeing their results from a BIA device, you may want to choose this method.

Output: Fat Mass, Non-Fat Mass

Why is BIA less accurate than DEXA?

BIA algorithms are derived using outdated methods, like hydrostatic weighing (commonly known as the Dunk Tank). The equations used to calculate body fat% using Hydrostatic weighing assume the density of bone and the density of muscle are the same for all people. But modern research has shown that not to be true. DEXA actually measures bone density and therefore is a more accurate method for measuring body fat. And as a result, you get higher body fat methods.

I thought you measured fat, not predicted?

No device actually measures fat, other than perhaps an MRI or through an autopsy. We don't recommend either. So when a company says they measure fat, they are misleading you. All devices commercially available are predicting fat through some algorithm. Styku's launch of BodyComp.AI presents the first ever attempt at (1) Changing the narrative in the industry to help people provide more accurate education to their customer, and (2) Give you the option to choose which model is best for your customers, and (3) Use artificial intelligence to improve the accuracy of the models.

To learn more, please visit our website to see the ground-breaking academic and scientific research behind our methods:

Styku Phoenix Research Paper