Spatial thinking has traditionally received much less educational focus in the curriculum, with most pedagogical content being related to, reading, mathematics and science (National Research Council, 2006). Spatial ability is correlated with performance in several different cognitive skills including maths (Lowrie et al., 2017) and reading (Giovagnoli et al., 2016). Furthermore, spatial abilities have been shown to respond well to focussed interventions (Uttal, Meadow, et al., 2013). Academic self-concept is an evaluative self-perception that is generated through the interpretation of, and experience of academic activities in a school environment (Marsh & Craven, 2006). A ten-year longitudinal study found that academic self-concept predicted educational attainment over and above prior achievement, even when controlling for other factors including family socio-economic status, family structure and academic achievement (Guay et al., 2004). This research project investigates what impact an intervention aimed at improving spatial skills can have on maths attainment and self-concept. In Phase 1, participants were administered an adapted version of the Self-Description Questionnaire-1 (SDQI) (Marsh, 1990). The SDQI measures self-concept in three academic areas (reading, mathematics, and general school). Participants were in years 4/5, in line with the appropriate age groups for the original SDQI. Additionally, I added a fourth academic area to the measure – spatial ability. Pupils (N=178) were administered a maths self-concept, reading self-concept scale and a spatial self-concept scale. To ensure understanding, participants were shown a novel 3-min video illustrating different spatial skills. Initial data analysis showed that the novel spatial self-concept measure has good internal consistency (Cronbach’s alpha=.845). Test-rest reliability scores were also good. Spatial self-concept is significantly correlated with both maths self-concept (r=.435, p<.01) and reading self-concept (r=.318, p<.01). Phase 2 of the study investigated the impact of a spatial skills training programme on maths attainment and spatial self-concept. The training programme consisted of 10, 20-minute sessions which focused on different spatial sub-skills. The spatial sub-skills selected were informed by Lowrie et al. (2017). This phase was a quasi-experimental design, with participants (N=57) recruited from two local schools. Participants completed pre- and post- intervention measures of maths attainment, and reading, maths and spatial self-concept. A process evaluation was also completed. Data analysis revealed a significant interaction between intervention group and time, in regards to maths attainment (F (1, 56) = 16.291, p>0.01). This demonstrates that engagement in the training programme resulted in an increase in spatial maths attainment. However, there was no impact upon any self-concept (spatial, maths or reading) observed in the statistical data. 11 Participants who engaged in the research programme were interviewed post-intervention. Interviews were transcribed and analysed using reflexive thematic analysis (Braun & Clarke, 2006). Three key themes emerged from the data: transfer, the dynamic nature of spatial skills, and new self-perceptions. This project has implications for the greater use of spatial skills training programmes for intervention for children with maths difficulties, which are discussed. Similarly, implications for how the results can inform future maths curriculum development are also evaluated. Finally the strengths and limitations of the research project and design are summarised.