This research addresses the challenge of pain assessment in individuals with cognitive deterioration (CD), a common aspect of aging and various neurological conditions. Due to difficulties in self-reporting, especially in severe cases, accurate pain diagnosis and management are hindered. The study explores the use of electroencephalography (EEG) and machine learning techniques to objectively measure pain in CD patients. Utilizing a BIS device, the research aims to identify EEG markers associated with pain, comparing them with an objective PANAID scale. The study targets patients in surgical departments, providing valuable insights into enhancing pain assessment for those unable to express pain through traditional subjective scales.

Cognitive deterioration (CD) may develop during the aging process and is a characteristic feature of various neurological and neurodegenerative diseases. Individuals with CD often face significant, prolonged, and intricate healthcare needs, frequently involving pain. However, effectively communicating pain characteristics becomes a challenge for individuals with CD, presenting a substantial obstacle to the accurate diagnosis and treatment of pain. CD affects various patient groups, although current data predominantly focus on dementia patients, revealing pain prevalence ranging from 40% to over 80%, depending on the context .

Due to its subjective nature, pain assessment relies predominantly on self-reporting. Individuals with CD often encounter difficulties in verbally expressing their pain due to limited intellectual and communicative abilities. Even when verbal skills are present, they may not guarantee valid pain reports. Consequently, pain assessment poses challenges for individuals with CD, particularly those with severe CD, elevating the risk of delayed or inaccurate pain diagnoses. Self-assessments or patient-reported measures are considered the gold standard in clinical pain assessment.

For individuals with compromised cognitive or linguistic abilities, or when self-assessment is impractical or invalid, behavioral measures can be employed. These tools capture facial expressions, vocalizations, or body movements as indicators of pain from an external observer's perspective, such as nurses, physicians, or healthcare providers. However, these parameters rely entirely on others being attentive to non-verbal pain signals, presenting a challenge as trained observers must reliably distinguish pain from various other facial and bodily expressions.

Developing objective measures reflecting the presence of painful states appears crucial to improving pain management in various clinical situations. In this regard, electroencephalographic (EEG) activation has been described as a cortical correlate of pain processing. Encouraging results have led researchers to consider increased gamma band activity as a potential indicator of pain presence applicable in clinical conditions.

This study employs a commonly used BIS device in hospitals to objectively measure pain levels in subjects with cognitive deterioration. Quantitative electroencephalography (qEEG) data will be obtained, and machine learning techniques will be applied for data analysis. Thirty patients experiencing cognitive decline, admitted to the general surgery and orthopedics departments at Volterra Hospital for significant surgical interventions, will be enrolled in the study. Concurrently, pain will be assessed using an objective PANAID scale and, if applicable, the NRS. The study aims to identify electroencephalographic markers of pain through machine learning techniques and establish correlations with pain levels obtained from the use of both subjective and objective scales

Inclusion Criteria:

• Subjects exhibiting at least moderate cognitive impairment as assessed by the Pfeiffer scale.

Exclusion Criteria:

• lack of consent

Patients admitted to general surgery and orthopedics, departments at Volterra Hospital with cognitive deterioration, undergoing major surgery.