This study examines the sentiment–returns relationship in both stock (S&P500) and cryptocurrency (Bitcoin) markets. An explorative wavelet analysis evidences period of episodic interconnectedness across different data frequencies. Therefore, Transfer Entropy (ET) measures remark the relative statistical significance, frequently outperforming traditional (VAR) estimates. In particular, ET methods successfully identify the mediating role of sentiments in connecting the two different markets. Hence, it is discussed how the potential cryptocurrencies indirect linkage with real economy moves through market sentiments.