Sleeping on low glycogen turned the fat programme on overnight, and left the mitochondrial markers alone

What sleeping low is

Sleeping low means training hard in the evening then going to bed without replacing the carbohydrate, so the overnight hours and the next morning's session happen with muscle glycogen low. It is studied because it gives the muscle a low carbohydrate signal for many hours without asking the athlete to eat less over the day.

Drawn from the paper's design.

Why they ran it

The idea that training with low carbohydrate availability amplifies the adaptive signal was already established, but the practical versions either reduced training quality or reduced total energy intake, which athletes will not accept. So the authors tested periodised nutrition rather than restricted nutrition. Same energy, same carbohydrate, different timing.

Drawn from the paper's stated aim.

Seven cyclists completed two trials with isoenergetic diets that differed only in when the food arrived. In one, they ate 8 grams of carbohydrate per kilogram of body mass before an evening high intensity session and then slept without eating. In the other they ate 4 grams per kilogram before the session and another 4 grams before sleeping. The next morning both trials did two hours of cycling while overnight fasted.

The two trials genuinely differed in muscle fuel. Muscle glycogen was higher in the fed condition at every time point after the evening session, at P less than 0.001.

Both trials raised PGC1 alpha and PDK4 messenger RNA at P less than 0.01, and PDK4 rose more in the fasted condition at P less than 0.05. Resting phosphorylation of AMPK, p38 MAPK and acetyl CoA carboxylase on day 2 was greater in the fasted condition, at P less than 0.05.

Fat oxidation during the two hour ride was higher in the fasted condition at P equals 0.01, alongside increases in acetyl CoA carboxylase phosphorylation and carnitine palmitoyltransferase 1, and in the messenger RNA for CD36 and fatty acid binding protein 3, all at P less than 0.05.

Methylation on the gene promoters for COX4I1 and fatty acid binding protein 3 rose four hours after the ride in both trials. Methylation of the PPAR delta promoter rose only in the fasted trial.

The headline null is in the authors' own conclusion. Delaying the post exercise feeding failed to augment markers of mitochondrial biogenesis.

The numbers

Carbohydrate intake, both conditions	8 g/kg body mass, timing differed
Muscle glycogen after the evening session	higher in the fed condition at all time points, P less than 0.001
PDK4 messenger RNA	raised in both, greater in the fasted condition, P less than 0.05
Resting AMPK, p38 MAPK and acetyl CoA carboxylase phosphorylation on day 2	greater in the fasted condition, P less than 0.05
Fat oxidation during the 2 hour ride	higher in the fasted condition, P equals 0.01
CD36 and FABP3 messenger RNA	increased in the fasted condition, P less than 0.05
PPAR delta promoter methylation	increased only in the fasted condition
Markers of mitochondrial biogenesis	not augmented by delaying post exercise feeding

Why this might happen

Shown in this study The study measured this step directly.

Sleeping with muscle glycogen low left AMPK, p38 MAPK and acetyl CoA carboxylase more phosphorylated at rest the next morning, and downstream of those sensors the fat handling programme was up. More PDK4 message closing the gate on carbohydrate, more carnitine palmitoyltransferase 1, more message for the fat transporter, and more fat oxidised on the ride to match.

The authors also report shifts in DNA methylation on the promoters of adaptive genes that correspond inversely with their transcription, presented as a correspondence rather than a demonstrated cause. And they report what did not move. Delaying the feeding did not augment the markers of mitochondrial biogenesis.

Drawn from the paper's reported results and stated conclusions.

What this does not show

It does not show that sleeping low makes anyone faster. There is no performance outcome and no training study. This is two days of measurement, and the practice is normally sold on an adaptation that accumulates over weeks.
The mitochondrial claim is the one that did not hold. The reason athletes are told to train low is mitochondrial biogenesis, and that is the specific thing the authors report as not augmented. The fat handling side moved; the building side did not.
The methylation finding is a direction, not a mechanism. Promoter methylation moving in the opposite direction to transcription is consistent with a regulatory link, and this design cannot establish that one caused the other.
Seven trained male cyclists, two days. A small crossover over a very short window. Nothing here describes what repeated exposure over a training block does, in either direction.

Where this leaves us

The same switch that a week of fat eating throws can be thrown by one evening of timing, and it can be read the next morning in the signalling proteins, the transcripts and the promoter methylation. What is refined rather than confirmed is the reason usually given for doing it. The fat handling arm responded, the mitochondrial building arm did not.

Seven trained male cyclists over two days, with total energy and carbohydrate held constant and only the timing moved.

A training block long enough to measure mitochondrial content and performance rather than the markers for them, comparing the same timing split repeated over weeks.

Caveats worth holding

Seven cyclists, two trials, five biopsies each across two days.
No performance outcome and no training intervention, so nothing here speaks to whether the practice works over a block.
The comparison is one specific timing split. A different split, or a different session, could land elsewhere.
Fat oxidation was measured on a morning ride done overnight fasted in both trials, so the fasted state is common to both and only the preceding glycogen differs.
The full discussion was not openly reachable, so the mechanism section rests on the paper's reported results and stated conclusions.
Trained male cyclists only.