Abstract Background Resting state (rs)-fMRI studies have found abnormal connectivity between the thalamus and other areas in schizophrenia. rs-fMRI may not have the capacity to establish connections between the functional patterns and symptoms or deficits that characterize the illness, and thus, the clinical consequences of this dysconnectivity remain poorly understood. Task-based functional connectivity can capture activity of a particular cognitive process and has shown to better characterize individual differences related to cognitive traits than rs-fMRI (Greene, Gao, Scheinost, & Constable, 2018). Only a few studies have explored thalamo-cortical connectivity using task-based fMRI, which could potentially enhance the capacity to match aberrant connectivity patterns with schizophrenia features. The aim of this study is to compare thalamic connectivity in rs-fMRI and AX-CPT task-based and its clinical correlations in subjects during the early phase of the illness Methods 115 FEP and 78 healthy controls were recruited from the UCDavis Early Psychosis Program. FEP were assessed using BPRS, SANS and SAPS. All participants underwent an MRI scan that included 6 minutes of resting state and 4 runs (40 trials per run) of the AX-CPT task. A subject-specific mask of the bilateral thalamus was extracted using Freesurfer segmentation of the structural images to generate subject-specific seed regions for both rs-fMRI and task-based connectivity. From the latter, only images during Cue A and Cue B were selected, as during these events a proactive cognitive strategy is gathered. To obtain thalamic connectivity, time-series BOLD activity in the seed region was correlated with functional images (both rs-fMRI and task-based fMRI) previously preprocessed in CONN toolbox including slice-time correction, realignment, normalization to an EPI template, smoothing and scrubbing. Voxel-wise correlation maps were compared between groups and corrected for multiple comparison using a threshold of FDR < 0.05. All analysis included age, gender, number of volumes scrubbed and mean motion as covariates. Results Increased connectivity was observed in FEP compared to HC between the thalamic seed and somatosensory regions and temporal areas. In contrast, decreased connectivity was observed between the seed and prefrontal regions (frontal pole) and cerebellum. AX-CPT task-based Likewise, increased connectivity was also observed in FEP in the AX-CPT between the thalamic seed and somatosensory and temporal regions. Decreased connectivity was also observed in FEP between the seed and the anterior cingulate and cerebellum. Increased connectivity in FEP during rs-fMRI was significantly correlated with total SANS score (p = 0.03). Discussion Besides rs-fMRI and task-based fMRI involve different cognitive processes, both study paradigms show similar pattern of between-group differences in thalamic connectivity, which reinforces the validity and reliability of previous studies on functional connectivity in FEP and schizophrenia. Additionally, decreased connectivity between the thalamus and anterior cingulate during task was also observed, suggesting that task-based connectivity may additionally demonstrate between-group differences in regions related to specific cognitive processes. Future studies comparing time-varying dynamic connectivity in rs and task-based fMRI may help elucidate the time course of these changes in connectivity in FEP.