Neuro-feedback for treating tinnitus seem to be a promising approach to help people suffering from chronic tinnitus. The past studies on this approach suffered interindividual variability in their results. The investigators' current hypothesis is that the variability of the results is the consequences of two possible flaws: lack of appropriate patient selection and lack of support of the patient during the therapy. This study is aimed at testing these hypotheses.

Many individuals with tinnitus have abnormal oscillatory brain activity in their temporal areas (Weisz et al. (2005), Schlee et al. (2014)). Led by this finding, attempts to normalize such localized pathological activity by neuro-feedback techniques have been tested (Dohrmann et al. (2007), Gütenspenger et al. (2017)). These attempts highlighted interindividual variability that can be explained by lack of selection of the patient population and lack of guidance through the therapy. The present study is aimed at addressing these issues by choosing more selectively a patient subpopulation (tinnitus associated with moderate hearing loss) and by implementing a guidance interface during the treatment.

The therapy will consist of 10 neuro-feedback training sessions of 29 minutes over 5 weeks. Each session will be composed of 6 blocks of 3 min in which the patient will be incited to practise a specific cognitive strategy (mental exercise such as "think to a music you like") and resting state measurements.

Each patient who has been recruited to fit our inclusion and non-inclusion criteria will first go through a clinical assessment of his initial judgment criteria metrics. Then subjects will go through the 5-week training and then will be evaluated on the same criteria just after the end of the therapy and at 3 months after the end of it.

Inclusion Criteria :

• Age > 18
• Social security affiliation
• Permanent, non-fluctuation, high pitch, bilateral or assimilated, tinnitus
• Mean value of the VAS at least at 6 between the intensity and the disturbance criteria at the recruitment interview.
• THI score superior or equal to 40 at the recruitment interview
• Written consent to the protocol
• Associated hearing-loss characterized by :
• mean value of the hearing threshold loss for the 250, 500 and 1000 Hz thresholds strictly under 25 dB (deciBel).
• at least one hearing threshold among the 2000, 4000, 6000 and 8000 Hz with at least 30 dB of hearing loss.
• mean value of the hearing threshold loss for the 2000, 4000, 6000 and 8000 Hz thresholds strictly under 70 dB.

Exclusion Criteria:

• Subjects under legal protection (guardianship, trusteeship or judicial protection)
• Notable cognitive disability impeding to understanding or performing the cognitive tasks
• Meniere's disease, chronic or serous otitis, acoustic neuroma
• Pulsatile tinnitus, somatosensorial tinnitus
• Epilepsy
• No antidepressant or anti-epileptic drug treatment
• DET (DETresse) questionnaire (measure of tinnitus associated distress) above or equal to 40
• Use of a sound generator during the therapy
• Pregnancy or breastfeeding
• Inability to wear the electrodes headset of the neuro-feedback device