Respiratory arousal is one of the most sensitive indices of autonomic activation, but despite close links between emotions and respiration in humans, few studies have assessed these links in animals. We have earlier reported that: a) respiratory activation is much more sensitive to brief alerting stimuli than traditionally used cardiovascular indices (1); b) these respiratory responses are sensitive to anxiolytics (2); and c) they are abolished after pharmacological inhibition of dorsomedial hypothalamus (3), a major neuronal centre of autonomic integration. Our current aim was to investigate potential roles of central amygdala (CAm) and medial prefrontal cortex (mPFC) in mediation of respiratory responses to alerting stimuli and stress. Outbred Wistar rats were implanted with guide cannulas targeting CAm (n=8) or mPFC (n=6) under isoflurane (2% in O2) anaesthesia. Respiratory recordings (whole-body plethysmography) were performed after microinjections of either GABAA agonist muscimol (200 nmol/ 200 nl) or saline to the targeted areas. Subsequently, rats were presented with six acoustic stimuli (40-90 dB, white noise, 0.5 sec) and subjected to restraint stress (15 min). Values presented are means ± S.E.M. compared by 1-tail paired samples t-tests between saline and muscimol pre-treatment trials. Acoustic stimuli produced brief increases in respiratory rate proportional to stimulus intensity, ranging from 12±9 cpm increase over baseline after 40 dB stimulus to 276±67 cpm increase after 90 dB stimulus. Inhibition of CAm by muscimol significantly decreased amplitudes of responses to 70, 80 and 90 dB stimuli (from 180±49 to 15±8 msec, p=.009; from 219±49 to 59±27 msec, p=.021; and from 276±67 to 103±49 msec, p=.043, respectively), but did not affect latencies of responses. Inhibition of mPFC, on the other hand, did not affect amplitudes, but increased latencies of respiratory responses to 80 and 90 dB stimuli (from 101±31 to 149±20 msec, p=.014, and from 33±4 to 81±16 msec, p=.024, respectively). Restraint stress elevated resting respiratory rate from 98±12 to 155±7 cpm (p<.001). Inhibition of CAm significantly reduced this stress-provoked increase in respiratory rate during the first 5 minutes of restraint from 57±12 to 12±15 cpm (p=.024), while inhibition of mPFC significantly decreased it during the last 5 minutes from 40±10 to 11±4 cpm (p=.02). Our study is the first where latency of autonomic responses to sensory stimuli was assessed at the timeframe compared to that of initial sensory processing in the brain. Our major novel finding is that PFC provides tonic influences favoring faster processing. Our data also indicate that integrity of amygdala neurons is essential for the expression of respiratory responses to brief arousing stimuli and for generating of initial, but not late-stage, tachypnoeic response to prolonged stressor.