Short-term cognitive loading deteriorates breathing pattern and gas exchange in adult patients with congenital central hypoventilation syndrome

Question

Human PHOX2B mutations result in life-threatening sleep-related hypoventilation (congenital central hypoventilation syndrome, CCHS). Most patients retain ventilatory activity when awake through a respiratory-related cortical network. We hypothesised that this need to mobilise cortical resources to breathe would lead to breathing-cognition interferences during cognitive loading.

Patients and methods

Seven adult CCHS patients (five women; median age 21) performed standard neuropsychological tests (paced auditory serial addition test - calculation capacity, working memory, sustained and divided attention; trail making test - visuospatial exploration capacity, cognitive processing speed, attentional flexibility; Corsi block-tapping test - visuospatial memory, short-term memory, working memory) during unassisted breathing and under ventilatory support. Ventilatory variables and transcutaneous haemoglobin oxygen saturation were recorded. Cortical connectivity changes between unassisted breathing and ventilatory support were assessed using electroencephalographic recordings (EEG).

Results

Baseline performances were lower than expected in individuals of this age. During unassisted breathing, cognitive loading coincided with increased breathing variability, and decreases in oxygen saturation inversely correlated with an increasing number of apnoeic cycles per minute (rho -0.46, 95% CI -0.76 to -0.06, p=0.01). During ventilatory support, cognitive tasks did not disrupt breathing pattern and were not associated with decreased oxygen saturation. Ventilatory support was associated with changes in EEG cortical connectivity but not with improved test performances.

Conclusions

Acute cognitive loads induce oxygen desaturation in adult CCHS patients during unassisted breathing, but not under ventilatory support. This justifies considering the use of ventilatory support during mental tasks in CCHS patients to avoid repeated episodes of hypoxia.

Impact of inspiratory threshold loading on brain activity and cognitive performances in healthy humans

In healthy humans, inspiratory threshold loading deteriorates cognitive performances. This can result from motor-cognitive interference (activation of motor respiratory-related cortical networks vs. executive resources allocation), sensory-cognitive interference (dyspnea vs. shift in attentional focus), or both. We hypothesized that inspiratory loading would concomitantly induce dyspnea, activate motor respiratory-related cortical networks, and deteriorate cognitive performance. We reasoned that a concomitant activation of cortical networks and cognitive deterioration would be compatible with motor-cognitive interference, particularly in case of a predominant alteration of executive cognitive performances. Symmetrically, we reasoned that a predominant alteration of attention-depending performances would suggest sensory-cognitive interference. Twenty-five volunteers (12 men; 19.5-51.5 years) performed the Paced Auditory Serial Addition test (PASAT-A and B; calculation capacity, working memory, attention), the Trail Making Test (TMT-A, visuospatial exploration capacity; TMT-B, visuospatial exploration capacity and attention), and the Corsi block-tapping test (visuospatial memory, short-term and working memory) during unloaded breathing and inspiratory threshold loading in random order. Loading consistently induced dyspnea and respiratory-related brain activation. It was associated with deteriorations inPASAT A (52 [45.5;55.5] (median [interquartile range]) to 48 [41;54.5], p=0.01), PASAT B (55 [47.5;58] to 51 [44.5;57.5], p=0.01), and TMT B (44s [36;54.5] to 53s [42;64], p=0.01), but did not affect TMT-A and Corsi. The concomitance of cortical activation and cognitive performance deterioration is compatible with competition for cortical resources (motor-cognitive interference), while the profile of cognitive impairment (PASAT and TMT-B but not TMT-A and Corsi) is compatible with a contribution of attentional distraction (sensory-cognitive interference). Both mechanisms are therefore likely at play.