Between 1974 and 1984 we have studied 204 control infants (C) comparing them with 650 SIDS siblings (SS) and 146 near-miss for SIDS (NM). These 1,000 full-term infants were recorded by day polysomnography (DPSG; n = 417), night polysomnography (NPSG; n = 257) and cardiopneumography (CPG; n = 2,600). Records were visually analyzed. In DPSG and NPSG, total amount of central, mixed and obstructive apnea as well as the percentage of periodic breathing was studied in each sleep state (active sleep, AS; quiet sleep, QS; indeterminate sleep, IS, and total sleep, TS) and over the total recording time (TRT). In CPG, only the total amount of central apnea and percentage of periodic breathing over TRT were studied. Infants were grouped according to postnatal age: < 5, ≥ 5 to ≤ 13, and > 13 to ≤ 26 weeks. In each age group results were compared as follows: C vs. SS, C vs. NM, and SS vs. NM for each parameter studied. Before 5 weeks and after 13 weeks there was no significant difference between C and SS, C and NM, and SS and NM in DPSG and NPSG for all categories of central, mixed and obstructive apnea as well as the percentage of periodic breathing in different sleep states and over TRT. Similar results were obtained in CPG for all categories of central apnea and percentage of periodic breathing over TRT. Between 5 and 13 weeks, results were comparable with those given above, except for central apnea between 2 and 5 s in DPSG, being more numerous in both NM and SS compared with C in AS (p < 0.001 and < 0.01, respectively), and in TS (p < 0.01 and < 0.001, respectively). 146 NM infants were compared according to whether they subsequently had no NM event (n = 117) or one or more NM events (n = 29). No difference was found between these 2 groups (in DPSG, NPSG and CPG) which could allow the risk of a subsequent NM event to be predicted. The same study was done in SS: five of them subsequently had one or more NM events. No significant differences were found between SS with or without an NM event. Four SS died unexpectedly (3 diagnosed as SIDS at autopsy) with no significant abnormalities on their recordings. In any case, according to our results DPSG, NPSG and CPG recordings do not permit the risk for SIDS in any infant to be predicted.