A key function of our brain is to identify temporal structures in the environment and use them to form expectations. These expectations allow us to plan and organize our behavior towards changes in the environment, and optimize the use of our attentional and motor resources. They also allow us to establish harmonious social interactions, coordinating us with our interlocutor during an exchange. Our ability to form temporal expectations seems to emerge very early but the development of this process is unknown. We only know that a set of basic skills, probably related to that ability, are present from birth. This suggests that temporal processing capabilities emerge during the prenatal period, but this has not been studied. The objective of this project is to study preterm infants brain's ability to process the intervals between stimuli and to form expectations on that basis.

40 patients (20 per group) : Presence of a functional response to stimulus omissions, only in the test group (regular presentation).

Group 1 will receive the test stimulation sequence (regular), Group 2 will receive the control stimulation sequence (random).

During the test sequence, a vibration will be repeated for 3 s interspersed with intervals of 5 s, creating an expectation about the presentation of the next stimulus. Every 8 to 12 vibrations (random), a vibration will be omitted. According to our hypothesis, we should observe an activity in the somatosensory cortex during the omission. A total of 10 omissions will be presented, bringing the total duration of stimulus presentation to 13 minutes.

During the control sequence, the interstimulus intervals are irregular, 3 to 7 seconds, so as not to induce expectations. Omissions in this case should not be associated with a functional brain response.

The presence of an activation in the primary somatosensory cortex will be assessed using Diffuse Correlation Spectroscopy (DCS), on newborns of 32, 33 and 35 weeks of corrected gestational age, during omissions in a regular vibrotactile stimulation sequence (test group) with respect to an irregular sequence (control group). A difference in activation during omissions between groups will reflect the presence of a processing of time intervals at the cortical level.

We will also assess the difference in amplitude and time to peak of the response between the corrected gestational ages and vigilance states.

Main endpoint: amplitude of the neurovascular response significantly higher and / or time to peak significantly shorter during the omissions in the test group (regular sequence) compared to the control group (irregular sequence), evaluated twice: at ages corrected SA 33 ± 3 days and 35 weeks ± 3 days , and if possible SA 32 ± 3 days (optional measure depending on the gestational age at birth and time to obtain consent).

Secondary endpoints: significant difference in amplitude and / or time to peak of the neurovasculat response between corrected gestational ages and vigilance states.

Inclusion Criteria:

• 40 premature infants born to SA 31 + 0 days to 32 weeks + 6 days , hospitalized in the neonatal unit of the University Hospital of Caen

Exclusion Criteria:

• Suspected neurological disorder (grade 3 or 4 intraventricular hemorrhage, periventricular leukomalacia), or motor condition that might interfere with tactile perception on the hand
• Intubation or CPAP
• Viral of bacterial infection
• Sedation at the time of measurement
• Parent(s) who are minor or unable to give free and informed consent
• No consent from the parents
• Non-inclusion request by the child's neonatologist