This paper introduces an information-based functional brain mapping approach, which contrasts with the traditional activation-based method. The activation-based approach focuses on identifying brain regions whose average activity changes across experimental conditions. In contrast, the information-based approach asks where in the brain the activity pattern contains information about the experimental condition. This is achieved by scanning the imaged volume with a "searchlight," whose contents are analyzed multivariately at each location in the brain. The traditional activation-based approach assumes that functional regions are activated as a whole, leading to the use of spatial smoothing to enhance detection of extended activations. However, this smoothing can obscure fine-scale activity patterns that may contain important neuroscientific information. The information-based approach, on the other hand, uses multivariate statistics to detect changes in activity patterns, which can capture both changes in spatial-average activity levels and changes in the activity pattern itself. The paper demonstrates that information-based mapping is more sensitive to focally distributed effects with equal power in all spatial-frequency bands than activation-based mapping. It also shows that such effects actually occur in fMRI data. Information-based mapping is particularly useful when the goal is to find where in the brain the regional spatial activity pattern differs across experimental conditions. The information-based approach is applicable to any fMRI experiment and can be used to detect differences in the spatiotemporal pattern of activity and differences regarding aspects of the pattern distribution other than its centroid. However, broadening the focus of an analysis to a more general class of effects comes at a cost in sensitivity. The paper proposes a method that extends the classical activation-based approach to detect distributed representations, which is more sensitive to weak extended activations. The paper also discusses the use of a spherical searchlight and the importance of the searchlight radius in determining the sensitivity of the mapping. It shows that a 4-mm radius searchlight yields near-optimal performance for both small and large regions of realistic shape. The paper also discusses the use of group analysis to address the challenge of spatial reference frames relating locations in different individual brains, which has much lower spatial precision. The paper concludes that information-based functional brain mapping is a promising alternative to the traditional activation-based approach, particularly for identifying regions containing information about the experimental condition. It is especially useful for experiments targeting distributed representations and for analyzing fMRI data at high spatial resolution.This paper introduces an information-based functional brain mapping approach, which contrasts with the traditional activation-based method. The activation-based approach focuses on identifying brain regions whose average activity changes across experimental conditions. In contrast, the information-based approach asks where in the brain the activity pattern contains information about the experimental condition. This is achieved by scanning the imaged volume with a "searchlight," whose contents are analyzed multivariately at each location in the brain. The traditional activation-based approach assumes that functional regions are activated as a whole, leading to the use of spatial smoothing to enhance detection of extended activations. However, this smoothing can obscure fine-scale activity patterns that may contain important neuroscientific information. The information-based approach, on the other hand, uses multivariate statistics to detect changes in activity patterns, which can capture both changes in spatial-average activity levels and changes in the activity pattern itself. The paper demonstrates that information-based mapping is more sensitive to focally distributed effects with equal power in all spatial-frequency bands than activation-based mapping. It also shows that such effects actually occur in fMRI data. Information-based mapping is particularly useful when the goal is to find where in the brain the regional spatial activity pattern differs across experimental conditions. The information-based approach is applicable to any fMRI experiment and can be used to detect differences in the spatiotemporal pattern of activity and differences regarding aspects of the pattern distribution other than its centroid. However, broadening the focus of an analysis to a more general class of effects comes at a cost in sensitivity. The paper proposes a method that extends the classical activation-based approach to detect distributed representations, which is more sensitive to weak extended activations. The paper also discusses the use of a spherical searchlight and the importance of the searchlight radius in determining the sensitivity of the mapping. It shows that a 4-mm radius searchlight yields near-optimal performance for both small and large regions of realistic shape. The paper also discusses the use of group analysis to address the challenge of spatial reference frames relating locations in different individual brains, which has much lower spatial precision. The paper concludes that information-based functional brain mapping is a promising alternative to the traditional activation-based approach, particularly for identifying regions containing information about the experimental condition. It is especially useful for experiments targeting distributed representations and for analyzing fMRI data at high spatial resolution.