<!-- Machine-readable version of https://dorsi.ai/blog/gluteus-medius-exercises-side-glutes. noindex. -->
# Gluteus Medius Exercises: Training the Side of the Hip

> Updated: 2026-07-18 · Source: https://dorsi.ai/blog/gluteus-medius-exercises-side-glutes

In one trial, squat and hip thrust training grew the gluteus maximus but not the smaller hip muscles. What a modelling study estimated about loading them.

![An illustration of a person holding a side plank on a dark green mat, seen from the side in a bright, near-empty room.](/blog/gluteus-medius-exercises-side-glutes.webp)

Most glute training is aimed at the gluteus maximus. The gluteus medius and minimus sit higher and more to the side, and they do a different job: holding the pelvis level when you are standing on one leg, which covers most of walking and all of running.

The evidence on training them directly is thinner than the volume of content about them suggests, and it is worth being clear about what kind of evidence it is.

<div class="takeaways">
<h2>Key Takeaways</h2>
<ul>
<li>In one 9-week trial, combined gluteus medius and minimus cross-sectional area at a single MRI slice showed little to no change in either the squat or hip thrust group, while gluteus maximus grew in both.</li>
<li>A modelling study estimated muscle forces across eight hip exercises in 14 female footballers. The highest-force tier for gluteus medius was the bodyweight side plank, loaded single-leg squat and loaded single-leg Romanian deadlift.</li>
<li>In that same acute protocol, adding 12-repetition-maximum resistance raised estimated peak gluteal forces by 28 to 150 N over the bodyweight versions. No study here followed anyone long enough to measure growth from these exercises.</li>
<li>"Dead butt syndrome" is a popular term rather than a diagnosis. The research here is about force and activation, not a condition.</li>
</ul>
</div>

## What one trial found about the smaller hip muscles

The [2023 Auburn trial](https://pmc.ncbi.nlm.nih.gov/articles/PMC10593473/) that compared nine weeks of hip thrust training against back squat training measured more than the gluteus maximus. It also measured the gluteus medius and minimus by MRI.

The measurement was combined gluteus medius plus minimus cross-sectional area taken at a single MRI slice, and it showed little to no change in either group, with small differences between them. The same nine weeks grew the gluteus maximus in both.

Several limits sit on that. It is one trial in untrained participants over nine weeks, testing two exercises rather than compound training in general, and a single-slice combined measurement cannot tell you what happened to each muscle separately or along its length.

A common explanation is that squats and hip thrusts load the hip front to back while these muscles mainly abduct and stabilise sideways. That is a reasonable hypothesis rather than something this trial tested, and it sits awkwardly with the modelling study below, which put loaded single-leg squats and Romanian deadlifts among the higher-force options for the medius.

## What a modelling study estimated

The most informative study here took a different approach to the usual activation ranking. A [2023 paper in *Medicine and Science in Sports and Exercise*](https://pubmed.ncbi.nlm.nih.gov/36918403/) combined motion capture, ground reaction forces and EMG as inputs to a musculoskeletal model, and used it to estimate forces in gluteus maximus, medius and minimus across eight hip-focused exercises. Fourteen female footballers each performed all eight, with and without 12-repetition-maximum resistance.

For the gluteus medius, the top tier was the bodyweight side plank, the loaded single-leg squat and the loaded single-leg Romanian deadlift. For the gluteus minimus, the top tier was the loaded single-leg Romanian deadlift and the bodyweight side plank.

The side plank reaching the top tier for both muscles without external load is the least intuitive part of the ranking, and it fits the idea that an isometric hold can demand high force from a muscle whose job is resisting movement.

The design limits matter as much as the ranking. These are modelled estimates, not measurements, since you cannot put a force transducer on a living gluteus medius. The model uses generic scaling rather than each participant's anatomy, and it was not validated against directly measured forces. The participants were 14 trained female footballers aged 18 to 32 performing single sessions. Nothing here followed anyone over time, so none of it establishes that these exercises grow the muscle or that a higher estimated force produces a better long-term result.

## What adding load did to the estimates

Within that same acute protocol, performing the exercises with 12-repetition-maximum resistance instead of bodyweight raised estimated peak gluteal forces by 28 to 150 N.

That is a within-session comparison of modelled forces, not a demonstration that loading produces more growth than choosing a different exercise would. It is suggestive, and it is the kind of claim that would need a training study to establish.

The banded side step was one of the eight exercises tested and did not reach the top tier for any of the three muscles. That is worth knowing if band work is the whole of your hip abduction training, though again it concerns estimated force in one session rather than what happens over months.

## About "dead butt syndrome"

The term gets used for gluteal weakness or inhibition associated with prolonged sitting, and it has become popular enough to be worth addressing directly.

It is not a clinical diagnosis, and neither study cited here investigated it. What this article can support is narrower: in one trial two common compound lifts did not grow these muscles much, and in one modelling study certain single-leg and side-lying positions produced higher estimated forces than others.

Hip or knee pain, a limp, or a leg that gives way is a clinical question and belongs with a physiotherapist rather than an article.

## A workable approach

Nobody has run the training study that would justify a confident prescription here, so what follows is inference rather than a finding.

The trial suggests squats and hip thrusts alone may not do much for these muscles. The modelling study suggests side planks, single-leg Romanian deadlifts and single-leg squats place the highest estimated demand on them, and that adding load raises that demand. Including one or two of those movements is a reasonable bet on that basis. It is a bet, not a result.

## References

1. Plotkin DL, Rodas MA, Vigotsky AD, et al. [Hip thrust and back squat training elicit similar gluteus muscle hypertrophy and transfer similarly to the deadlift](https://pmc.ncbi.nlm.nih.gov/articles/PMC10593473/). *Frontiers in Physiology*. 2023;14:1279170. doi:10.3389/fphys.2023.1279170. Free full text.
2. Collings TJ, Bourne MN, Barrett RS, et al. [Gluteal Muscle Forces during Hip-Focused Injury Prevention and Rehabilitation Exercises](https://pubmed.ncbi.nlm.nih.gov/36918403/). *Medicine and Science in Sports and Exercise*. 2023;55(4):650–660. doi:10.1249/MSS.0000000000003091.