A destructive read is the most common type of test reaction in memory cell circuitry. Information is stored in a memory cell by placing objects in known positions, or by changing the state of a stable or periodic toggle circuit. A destructive-read test consists of sending one or more signals to the memory cell. A distinct output signal is produced for each possible state of the memory cell, which is reset to a known "zero" or "rest" state; see for example boat-bit, keeper, and demultiplexer.
To permanently store information in a destructive-read memory cell, the output signal(s) must be used, in part, to send appropriate signals back to the memory cell to restore its state to its previous value. With output looped back to input, this larger composite circuit then effectively becomes a non-destructive read memory cell.