Table 3.

Exploring the predictive relationship between social media usage and depression.

Author	Main findings
Hartanto et al.151	Social media usage increases among depressive individuals
Figueredo et al.78	Semantic mapping of emoticons improves the performance.
Stankevich et al.79	Future work needs to involve applying semantic role labeling to obtain better results.
Lara et al.77	DeepBoSE outperforms conventional Bag-of-Features(BoF) representations.
Hussain et al.152	Proposed depression lexicons that distinguish depressive individuals.
Ramiandrisoa et al.86	Analyzing users’ social signals could be considered for further analysis.
Liaw et al.153	Topic modeling features such as liked tweets can be useful.
Guo et al.84	Fused the lexical features using a correlation-based metric to enhance prediction effectiveness.
Cui et al.83	Capture deep emotional information from the input embeddings with a pre-trained TextCNN.
Zogan et al.87	The model captures semantic features from user timelines for depression detection.
Tlatelpa et al.80	User characteristics and sentiment analysis improved depression detection performance.
Cha et al.82	Proposed lexicon features for depression detection.
Primack et al.154	Using multiple social media platforms is associated with depression.
Primack et al.155	Social media use is associated with the development of depression.
Vedula et al.156	Depressed users exhibit reduced online activities, increased negative sentiment, and self-focused pronoun usage.
Nesi et al.157	More frequent negative emotional reactions to social media are linked to more severe depression symptoms, especially among female subjects.
Thorisdottir et al.158	Time spent on social media has a stronger relationship with emotional distress among female subjects.
Ghosh et al.159	Depressed users frequently use negative words and mostly post late at night, in addition to increased use of personal pronouns and sharing personal events.
Aragon et al.160	Representations based on fine-grained emotions can more comprehensively capture users experiencing depression.
Puukko et al.161	Depressed individuals increasingly use active social media during early and late adolescence.
Robinson et al.162	Depressed individuals are more likely to compare themselves to others and dislike being tagged in self-perceived unflattering pictures.
Choudhury et al.163	Social media can provide valuable indicators of depression onset, including decreased social activities, increased negative emotions, focus on personal and medical issues, and more frequent expressions of religious involvement.