Parallel Magnetic Resonance (MR) imaging is a technique that maintains or improves spatial resolution at reduced acquisition rates through the use of multiple receiver coils. Gains in temporal resolution are achieved by under sampling k-space in a process known as acceleration and the scan time reduction is proportional to the acceleration factor (R). The extent to which parallel imaging can be exploited is limited by the reduction in signal-to-noise ratio (SNR), proportional to 1/√R, that is a consequence of data under-sampling. To explore the application of parallel MR imaging to detection of breast cancer, a set of tissue-mimicking water and oil phantoms were imaged using a GE 3 T MR imaging system and dedicated 8-channel breast coil. Various one-dimensional acceleration factors were approximated from fully sampled data in an attempt to quantify temporal gains and the resultant SNR penalty from accelerated imaging. SNR degradation for acceleration factors of 1.84, 3.16 and 4.09 were found to be 70%, 57% and 31% of SNR obtained with a fully sampled data set (R=1) respectively in a central slice. SNR efficiency (mean SNR/√t) for each acceleration factor was determined to be 0.944, 0.872 and 0.626 respectively for the same acceleration factors as above, suggesting that trade off between scan time and SNR penalty can be maintained up to acceleration factors of approximately R=3. Thus, 8-channel parallel breast MR imaging at 3 T appears capable of achieving relatively high SNR at temporal resolutions approaching three times better than conventional breast MR imaging.