Sentiment analysis, a process aimed at identifying opinions and emotions present in textual data, presents significant challenges despite the progress made in text classification thanks to the new generation of neural language models such as Transformers. These challenges are particularly evident in situations characterized by limited lexical production or specific contexts with specific domains. One of the main challenges in sentiment analysis is the correct identification of negation. Detecting such relationships within Italian texts is crucial. Negation can lead to the opposite meaning of a sentence, significantly influencing overall understanding. In this approach, it is important to follow linguistic cues of negation such as adverbs or complex constructions and can be used in various contexts and languages, making it more flexible. In fields like bioinformatics and medical literature, negation detection plays a crucial role in building connections between genes and diseases, medication prescription history, or the lack of documented complaints by patients. The purpose of this article is to provide a general overview of the search for more effective solutions to detect negation in Italian texts, addressing current limitations such as the scarcity of datasets in the medical domain. Future work will focus on the use of high-level approaches to further improve the performance of negation identification and enhance the stability of sentiment analysis, especially in complex contexts.