Sepsis-associated encephalopathy (SAE), is one of the most common organ dysfunction during the acute phase in sepsis and septic shock. Electroencephalogram (EEG) and auditory evoked potentials (AEPs), which reflect different aspects of brain function, are the most commonly used neurophysiological indices to detect acute brain dysfunction in critically ill patients including sepsis and septic shock. AEPs show the systemic responsiveness of the central nervous to auditory stimuli, so they can be considered a direct measure of brain responsiveness. Mismatch negativity (MMN) is a change-specific component of ERPs, which elicited by a deviant stimulus occurring in a sequence of repetitive stimuli. This component is thought to represent the automatic and unconscious detection of acoustic changes which requires good perceptual discriminative capacity and iconic memory. The peaks of MMN appear at 100 ~ 250 ms from deviant stimulus onset; with increasing magnitude of stimulus change, the peak latency of MMN was shortened and the amplitude increased. Since MMN can be elicited even in the absence of attention, subjects do not need to actively participate. The MMN has been extensively demonstrated to be used in the prediction of awakening in comatose patients for various reasons, and also has been reported to predict awakening in deeply sedated critically ill patients recently. However, it remains unclear whether SAE affects MMN in amplitude and latency that reflects cognitive processing of the auditory information.

Patients with sepsis and septic shock who met the inclusion criteria were screened daily on the CAM-ICU scale, and those with positive CAM-ICU were diagnosed with SAE.All patients were tested for event-evoked potentials on Day 1 and Day 3 after inclusion and were followed up to Day 28 after discharge. The investigators intend to observe the dynamic change of MMN amplitude and latency between SAE and non-SAE groups. Logic regression analysis was used to determine whether the change of MMN was a predictor of SAE.

Patient characteristics (including age and sex), date and time of ICU admission, category of admission (medical, surgical or emergency), main reasons for admission, co-morbidities, complications and duration of sepsis before admission were collected. The severity of illness assessed by the acute physiology and chronic health evaluation scoring system (APACHE) II and sequential organ failure assessment (SOFA) score, bladder temperature, heart rate (HR), systolic blood pressure (SBP), diastolic blood pressure (DBP), respiratory rate (RR), the level of consciousness assessed by the GCS score, mechanical ventilation (MV) or not, the depth of sedation assessed by RASS, and dose of sedative drugs at 1 and 3 days after admission were also collected.The EEG signal was recorded on the 1st and 3rd day after admission using silver-silver chloride disc electrodes placed on the scalp according to the international 10-20 system. Four electrode locations (left frontal [F3], right frontal [F4], frontal midpoint [Fz] and central [Cz]) were used. All electrodes were referred to the both earlobe and the ground electrode on the forehead. The electrode-skin impedances were kept below 5 kΩ. The EEG signal was amplified and digitized continuously at 279 Hz using EMMA (ERP measuring machine; developed and custom-made in the Department of reinject Medical Equipment Co., Ltd, Guangzhou, China) Background EEG was recorded for 10 min during sleep and/or while the patients lay motionless with their eyes closed in a quiet room. Auditory stimulation was then set to 'on' so that ERPs could be recorded. The stimulation was applied according to an oddball paradigm, which consisted of 85% standard (800 Hz) and 15% deviant (560 Hz) stimuli, with an interstimulus interval of 1 s. The duration of each stimulus was 84 ms, including 7 ms rise and fall times. Altogether 600 stimuli were delivered through earphones to the right ear for each measurement, corresponding to a recording time of about 10 min. The stimulus intensity was set at 75 dB. According to the patient's original EEG, the investigators recorded the main components of the background wave, whether there was periodic discharge and low voltage.The amplitude and latency of MMN components of the event evoked potential.

Inclusion Criteria:

• ages between 18 and 80 years;
• expected stay in the ICU of > 72 h;
• patients diagnosed with sepsis or septic shock;
• informed consent was signed by the patient or relatives;

Exclusion Criteria:

• at terminal stage of disease;
• primary brain injury (such as traumatic brain injury, stroke, cardiac arrest, intracranial infection, epilepsy, Alzheimer's disease, Parkinson disease and meningitis etc.);
• acute mental deterioration secondary to non-septic metabolic disorders with organ dysfunction (hepatic encephalopathy, pulmonary encephalopathy, severe electrolyte imbalance, severe blood glucose disorders etc.);
• history of craniocerebral surgery;
• psychiatric illness;
• use of psychiatric medications;
• impaired hearing; participated in other clinical trial;
• pregnant or lactating women;
• expected death within 72 h after admission.

patients with confirmed sepsis or septic shock

It is not yet known if there will be a plan to make IPD available