22 December 2020

Running out of GAS – Viewing stress & adaptations in a different light


The GAS model has been taught to many as the way in which stressors are met and dealt with by the human body. This is particularly the case with regards to the training realm, as it is an easy and straightforward way of seeing how we respond to training stressors. But is it really that…

The GAS model has been taught to many as the way in which stressors are met and dealt with by the human body. This is particularly the case with regards to the training realm, as it is an easy and straightforward way of seeing how we respond to training stressors. But is it really that useful?

This article aims to highlight some of the flaws of this model and why we should look elsewhere in our bids to explain training stressors and the adaptations to those stressors. We’ll first begin with a little trip through history, before discussing why it might not be so useful as a principle. We’ll then take a look at some other, possibly more appropriate models  that can give us a more in depth understand of training, stress and the resulting adaptations.

In the 1930’s, Hans Selye brought forth the GAS (General Adaptation Syndrome) Model – which describes an organism’s generalized stress response to a variety of different stressors. The model was developed after examining how rodents responded after administration of varying levels of lethal drugs, extreme environmental conditions, starvation, and strenuous exercise (yeah, pretty gnarly stuff).

The research was primarily focused on how the immune and endocrine systems of mice would respond to these stressors. What they found was the response could be broken down into three main phases: Alarm -> Resistance -> Exhaustion. 

Stage 1: Alarm phase – An imposed stress/threat compromises status of organism

Stage 2: Resistance – A response occurs in an effort to keep organism within homeostasis (or beyond i.e. adaptation)

Stage 3: Exhaustion – Occurs when the magnitude of stress exceeds the “adaptive resources” available to cope with it. 

Seeing the model as somewhat useful, the Russians (it’s always the Russians) used it as a theoretical basis for training cycles shortly after studying the routines of successful and unsuccessful athletes at the 1956 Olympics. It was a way to organise training stress in a manner conducive to getting the best results (at least they thought it was).

The idea here was that the Alarm Phase would be the initial exposure to the training stressor, which would see an immediate drop in performance because of fatigue. The Resistance Phase would be a return to baseline and/or beyond as a response to the stressor, this would see an increase in performance. The Exhaustion Phase would be the inability to tolerate the sum of stressors, resulting in a big drop in performance or negative outcome, like injury, pain, etc.

Sounds reasonable, right? 

This line of thinking also brought rise to the Supercompensation Hypothesis i.e. the SRA curve (Stress, Recovery, Adaptation), which I’m sure you’ve heard about it in the past. This model is heavily influenced by the GAS paradigm.

This is commonly used to teach how the stress from training results in adaptations or improved fitness. It’s also been the basis of many periodisation models – we impose a stressor on ourselves (train), we recover (bring ourselves back to baseline), and granted all things go well, we can go beyond this imaginary baseline. If we fail to apply the stimulus within a given period, we lose those adaptations (principle of reversibility).

Now, just because the GAS model & SRA curve are popular and still used today, does it really mean it’s the most useful model with respect to training? According to Buckner et al.( 2017) this is actually a misapplication of the model. We are humans performing non-lethal doses of voluntary exercise, making the GAS Model unsuitable as a foundational principle when aiming to organise training stress. 

Note: Just thought I’d chime in to say that to be fair to Selye he never intended or thought of the GAS model to be used in training/exercise, others just used this concept and its popularity grew with coaches and trainers over time. 

Kiely (2018) discusses some other features of the paradigm that might make it unsuitable for an appropriate training model:

  • The GAS model looked only at acute responses to (extreme) stressors. Not chronic. So, it doesn’t tell us much about long term training responses.
  • Stress and the stress response are seen as independent of the brain. This suggests that there is no psycho-emotional influence on the stress response.
  • That the stress response is both non-specific & predictable
  • While Selye saw how the response occurred, he could never find the one biological source that would kickstart this response – which, he called the first mediator. Dr Mason in the 70’s found that the response to a stressor was psycho-emotional in nature (Mason, 1975). We’ll see why this is important later.

While the model has no doubt been useful and a driving force behind a lot of the research in regard to stress and adaptations, it does seem to fall short in many important areas. Those using the GAS model reduce adaptations to a simple physical input = physical output scenario, as if we were to exist in a biological vacuum. It also suggests that training responses are predictable and somewhat linear to the stressors imposed. Anyone that has been training for some time knows that even with the ‘perfect’ program, what we expect to happen isn’t always the case. It’s somewhat disinterested in the dynamism of individual responses. 

Even with these flaws, and our developing understanding of stress, the GAS model is still heavily used to explain training adaptations even to this day.  

In saying this, these flaws didn’t go unnoticed. These gaps in the model gave rise to other paradigm’s, where they could aim to fill in the missing pieces. Proposed by Banister in 1982, the Fitness-Fatigue model was one of the pieces. It grew quickly in popularity across multiple sporting and exercise realms as it was more specific to training and the training response. There was much more nuance behind how these acute and chronic changes come to be. 

We won’t be discussing this today, but we will touch on another area I think is very useful in helping us conceptualise stress and adaptations a little better than in previous models. It also answers some of the shortcomings of GAS. 

In 1988, Sterling and Eyer brought forth the concept of Allostasis, defined as the active process of maintaining homeostasis through the adaptive change of internal conditions to meet perceived and anticipated demands (Zsoldos & Ebmeier 2016). It’s the idea of “stability through change”. 

Bit of a mouthful, I know. John Kiely (2018) helps a little with this by describing it as “the complex set of emotional, physiological, immunological, and psychological processes that intimately collaborate to establish a new set of internal conditions best fitting current circumstances“. While not a specific training model, it allows us to see how stress and the stress response aren’t as general or predictable as older models of stress like GAS make it out to be. 

Sterling and Eyer (1988)

With the previous models, training is often reduced to mechanical input → physical/biological output. This assumes that the training variables we apply e.g. intensity and volume, are the sole dictators of what the training outcome will be. While they are certainly the main drivers of the training response, there are other variables we’ve discussed that will influence the degree of the outcome. 

The reason why the Allostasis model can be useful is that no longer considers the stress response as independent of the brain, but that it also has psycho-emotional influence, which no doubt plays a role in the subsequent adaptations as you can see below: 

(Kiely, 2018)

This opens the doors to training being more than just intensity or the amount of volume completed. Like I said previously, there’s no doubt that the training stimulus will be the primary instigator of a specific training response, butwe can also now see it as one main driver among many. 

With this update on how we see stress and adaptations, we can appreciate that fitness can no longer be viewed in the one-dimensional lens of purely the physical. There’s going to be many other factors involved that will shape and determine the end product.

(Kiely, 2018)

This model also feeds into the importance of ‘individualised’ programming and how while one program may work for one person, it may not work for others. 

Some studies are pretty useful in demonstrating just how individualised some of these training outcomes are when the program variables are the same (intensity, volume, etc.). In Erskine et al (2010) we see such wild variability in both strength and hypertrophy outcomes after 9 weeks of a particular training program:

Hubal et al. (2005) is another interesting study that found some incredible differences in results in 585 subjects. The range of results went from -2 to 49% in regard to muscle size across a 12-week program – yes, someone actually lost muscle size. 1RM strength testing (of the elbow flexors) ranged from no improvements in some participants to a whopping +250% increase. The 250% increase only equated to about 10kg, but still a wild difference among participants.

Another study by Timmons (2011), looks at variability and individual differences from a genetic perspective, finding differences in results across multiple training regimes i.e. endurance and hypertrophy training. He found that differences in gene sequences (polymorphisms) also contributed to some of this variability in training responses. 

What also contributes to this idea of individual responses is that when we look to the opposite end of the spectrum, when programs are individualised – we actually get somewhat similar results across the board (Coakley & Passfield 2018).

I think you get my point by now; we are exposed to so many different stressors and our responses to said stressors can vary enough that the resulting output may not be as predictable as we’d like to think. 

This has obvious implications down the line for how we implement certain training or even recovery strategies. 

The Allostatic model can be used to contribute to previous models, as it adds some pieces to the never-ending puzzle. This means making more informed choices with training and recovery while respecting the complexity of human systems.

To wrap it all up, I hope the above has let you zoom out a little to see the bigger picture. It is no longer enough to think about training as if it were to exist in its own bubble. It is to think beyond, and to build awareness of not only what you’re doing in the gym, but outside of it as well. It adds to the following: 

  • The stress response and in turn adaptations aren’t general or predictable, but unique to each individual.
  • We no longer just consider the physical stressors when training.
  • The totality of stressors should be considered before assigning the training stimulus. As this will determine if the outcome will be positive or maladaptive.

Your email address will not be published. Required fields are marked *

Send this to a friend