Patients with depressive disorder (DD) have disordered emotions (e.g., sadness, anhedonia), actions (e.g., retardation), and cognition behaviors (e.g., diminished ability to think or concentrate) compared to healthy individuals (McCarron, Shapiro, Rawles, & Luo, 2021; Roberts, 2019). In recent years, a new non-invasive technology functional magnetic resonance imaging (fMRI) has been introduced to study pathophysiology in patients with DD. Convergent studies suggested that the abnormal functioning of frontostriatal networks may contribute to anhedonia in patients with DD (Price & Drevets, 2010a, 2012b). Moreover, the low striatal dopamine levels in depression may fail to transform the viscerally charged information to cortices for contextual processing and reappraisal, thus leading to the salience of negative information (Hamilton et al., 2012). Other studies indicated that patients with DD have shown abnormal functional activation in the caudate nucleus when processing cognition tasks (Harms et al., 2019; Pan et al., 2013). As a part of the striatum, the altered functional activation of the caudate nucleus and its anatomical abnormalities are related to the behaviors of patients with DD.

The caudate nucleus is composed of a large anterior head, a body, and a thin tail that wraps anteriorly such that the caudate nucleus head and tail can be visible in the same coronal cut (Driscoll, Bollu, & Tadi, 2020). And behavioral domain resulted in cognition, emotion, and action-related structures and networks primarily localized in different parts of the caudate nucleus(Robinson et al., 2012). One study (Sacchet, Livermore, Iglesias, Glover, & Gotlib, 2015) revealed that the head of the caudate nucleus is connected with the lateral and medial prefrontal cortices and is involved in working memory and executive functioning. The body of the caudate nucleus receives input from the prefrontal cortex relates to both cognitional and emotional controls (Kotz, Anwander, Axer, & Knösche, 2013). The tail of the caudate nucleus is connected to the motor and premotor cortices which usually process visual information and control movement (Robinson et al., 2012).

Moreover, a previous study (Kumral, Evyapan, & Balkir, 1999) suggested that acute infarcts and hemorrhages of the caudate nucleus may cause serious depression. Whereafter, a postmortem study estimated neuronal density and volume, and glial density of cells from human brain tissue of the caudate nucleus, and the results suggested there were reductions in both the dorsolateral and ventromedial areas of the caudate nucleus in the DD group (Khundakar, Morris, Oakley, & Thomas, 2011). To further explore how the caudate nucleus activation changed during the active brain state scientists have been using task-based fMRI (t-fMRI) to examine regions of pathological brain activation in patients with DD during the completion of tasks by measuring blood oxygen level-dependent (BOLD) signal changes (Kannurpatti, Rypma, & Biswal, 2012) which are proposed as a potential biomarker for determining the degree of the impairment of the caudate nucleus (Ances et al., 2006).

Hammar et al. (2016) conducted an experiment that combined a Stroop task and an N-back task, in which they found a sample of remitted and partially remitted DD patients when contrast to healthy controls (HCs) exhibited normal activity in the dorsal lateral prefrontal cortex (DLPFC) and anterior cingulate cortex (ACC) but reduced activity in the striatum, particularly in the caudate nucleus. While during the Stop Signal Task (SST), DD participants demonstrated greater BOLD activation relative to the comparison group in predominantly left-lateralized fronto-striatal-limbic circuitry that included the left side of the caudate nucleus (Bobb et al., 2012). However, there have been no meta-analyses that are specifically conducted on the caudate nucleus using Activation Likelihood Estimation (ALE) (Eickhoff, Bzdok, Laird, Kurth, & Fox, 2012) which can summarize the statistical relation between study characteristics and findings.

Specifically, ALE models the activation foci as 3D Gaussian distribution centers at the reported coordinates and then calculates the overlap of these distributions across different experiments (ALE treats each contrast in a study as a separate experiment) (Caspers, Zilles, Laird, & Eickhoff, 2010). This allows experiments with larger samples to have greater statistical certainty in the meta-analysis. Thus, concordance between studies can be identified by creating statistical probability maps as a measure of the likelihood of morphological activation on a voxel-wise level across the entire set of studies entering the meta-analysis. ALE has been previously applied with success in DD populations (Ashworth, Brooks, & Schiöth, 2021; Zhang et al., 2022; Zheng, Van Drunen, & Egorova-Brumley, 2022). Therefore, ALE is considered to be the best way to test whether the hypo-/hyper- activation of the caudate nucleus is located in different parts of the caudate nucleus during t-fMRI.

Moreover, recent studies demonstrated patients with DD have shown abnormal activations of the caudate nucleus in different task categories (emotion (Bomyea et al., 2021; Tu, Chen, Yang, Galecki, & Su, 2020), cognition (Hammar et al., 2016), and affective cognition (Alders et al., 2020)). In the emotional behavior domain, one previous study (Silk et al., 2017) suggested hypoactive caudate in the onset of depression in adolescents when listening to previously recorded audio clips of their own mothers’ praise, criticism, and neutral comments compared with HCs. In the cognition behavior domain, patients with DD showed deactivation in the caudate nucleus compared to HCs when they suppressed their thoughts (Carew, Milne, Tatham, MacQueen, & Hall, 2013). During an Oddball task embedded with sad or neutral images that belong to the affective cognition domain, HCs showed greater activation in the caudate nucleus (Dichter, Felder, & Smoski, 2009). Perhaps, it is the different parts of the caudate nucleus (caudate head, caudate body, and caudate tail) that may functionally relate to different behavior domains (Kotz et al., 2013; Robinson et al., 2012). Therefore, we hypothesized that the different abnormal caudate activities may relate to the different behavior domains in patients with DD.

To investigate the abnormal caudate subregion activity and the relationships between the different abnormal caudate activity and different behavior domains in patients with DD. Meta-analyses were performed using the method of ALE with two different classification methods. One was conducted based on hypo-activation and hyper-activation of the caudate nucleus, respectively. Then according to the Brainmap behavioral domain (brainmap.org | Home) and the discipline of affective cognition (Elliott, Zahn, Deakin, & Anderson, 2011) based on different task-performing, the studies were divided into the emotion domain, cognition domain, and the affective cognition domain.