**1. Introduction**

Approximately 395,000 adults experience an out-of-hospital cardiac arrest (OHCA) annually in the US, and their overall survival rate is only 6–11% [1–3]. To prevent death or irreversible damage to vital organs, such as the brain, high-quality cardiopulmonary resuscitation (CPR) is necessary [4,5]. Multiple physiologic measurements have been suggested as indicators of the effectiveness of CPR. End-tidal carbon dioxide (ETCO2) is a widely-used indicator for the pulmonary circulation, and the ETCO2-directed feedback methods are reported to improve the likelihood of return of spontaneous circulation (ROSC) [6].

Recently, achieving good neurological recovery has been regarded as one of the major goals of CPR, because it can influence survivors' quality of life and their socioeconomic burden [7,8]. However, ETCO2 mainly reflects the systemic circulation, and thus is not adequate to monitor the cerebral circulation or physiological responses of the brain during CPR. Carotid blood flow, the blood supply to the brain, can reflect the cerebral circulation directly. However, its measurement requires an ultrasonic volume flow meter, as well as a skilled operator.

Non-invasive electroencephalography (EEG) can be an alternative to overcome these drawbacks. Portable and low-cost EEG headsets and sensors are currently available out-of-hospital [9]. EEG activity during CPR is reported to be sensitive to cerebral circulation [10,11]. Once cerebral oxygenation decreases due to cardiac arrest (CA), the EEG activity gradually enters the isoelectric state [12–14]. However, the EEG activity can return to the pre-arrest state when ROSC was achieved [15,16]. Effective CPR maintains a certain degree of cerebral electrical activity, changing the EEG activity from isoelectric status to large-amplitude and low-frequency status with bispectral index score (BIS) above 40 [17]. As an important tool for determining the prognosis of ischemic episodes of CA patients, EEG signal is routinely monitored for post-resuscitation treatment [18]. The application of EEG monitoring has expanded to the CPR situation, and distinctive EEG patterns are suggested as possible markers for the quality of cerebral resuscitation and oxygen delivery [19]. To date, however, the direct relationship between the carotid blood flow (CBF) recovery and the EEG during CPR has been rarely discussed.

In this study, we focused on the investigation of the relationship between the recovery of carotid blood flow and non-invasive EEG parameters, including time- and frequency-domain parameters, and entropy indices between defibrillation attempts. We applied a single-channel EEG measurement device that was developed in our laboratory and designed a ventricular fibrillation (VF) swine model with simultaneous measurements of EEG and hemodynamic data, including CBF. We hypothesized that CBF recovery may improve cerebral electrical activity, which can result in EEG changes, even during short intervals between defibrillation attempts.
