Work­ing mem­o­ry is a lim­it­ed resource: Brains can only main­tain small amounts of sen­so­ry input (mem­o­ry load) over a brief peri­od of time (mem­o­ry decay). The dynam­ics of slow neur­al oscil­la­tions as record­ed using mag­ne­to- and elec­troen­cephalog­ra­phy (M/EEG) pro­vide a win­dow into the neur­al mechan­ics of these lim­i­ta­tions. Espe­cial­ly oscil­la­tions in the alpha range (8–13 Hz) are a sen­si­tive mark­er for mem­o­ry load. More­over, accord­ing to cur­rent mod­els, the resul­tant work­ing mem­o­ry load is deter­mined by the rel­a­tive noise in the neur­al rep­re­sen­ta­tion of main­tained infor­ma­tion. The audi­to­ry domain allows mem­o­ry researchers to apply and test the con­cept of noise quite lit­er­al­ly: Employ­ing degrad­ed stim­u­lus acoustics increas­es mem­o­ry load and, at the same time, al- lows assess­ing the cog­ni­tive resources required to process speech in noise in an eco­log­i­cal­ly valid and clin­i­cal­ly rel­e­vant way. The present review first sum­ma­rizes recent find­ings on neur­al oscil­la­tions, espe­cial­ly alpha pow­er, and how they reflect mem­o­ry load and mem­o­ry decay in audi­to­ry work­ing mem­o­ry. The focus is specif­i­cal­ly on mem­o­ry load result­ing from acoustic degra­da­tion. These find­ings are then con­trast­ed with con­tex­tu­al fac­tors that ben­e­fit neur­al as well as behav­ioral mark­ers of mem­o­ry per­for­mance, by reduc­ing rep­re­sen­ta­tion­al noise. We end on dis­cussing the func­tion­al role of alpha pow­er in audi­to­ry work­ing mem­o­ry and sug­gest exten­sions of the cur­rent method­olog­i­cal toolkit.