This study presents a time-dependent stochastic user equilibrium (TDSUE) traffic assignment model within a probit-based path choice decision framework that explicitly takes into account temporal and spatial correlation (traveler interactions) in travel disutilities across a set of paths. The TDSUE problem, which aims to find time-dependent SUE path flows, is formulated as a fixed-point problem and solved by a simulation-based method of successive averages algorithm. A mesoscopic traffic simulator is employed to determine (experienced) time-dependent travel disutilities. A time-dependent shortest-path algorithm is applied to generate new paths and augment a grand path set. Two vehicle-based implementation techniques are proposed and compared in order to show their impact on solution quality and computational efficiency. One uses the classical Monte Carlo simulation approach to explicitly compute path choice probabilities, and the other determines probabilities by sampling vehicles' path travel costs from an assumed perception error distribution (also using a Monte Carlo simulation process). Moreover, two types of variance-covariance error structures are discussed: one considers temporal and spatial path choice correlation (due to path overlapping) in terms of aggregated path travel times, and the other uses experienced (or empirical) path travel times from a sample of individual vehicle trajectories. A set of numerical experiments are conducted to investigate the convergence pattern of the solution algorithms and to examine the impact of temporal and spatial correlation on path choice behavior.