Keep it familiar, keep it instrumental, and keep the one you already use. The obvious diagnosis is that you got used to it, but the research on habituation points both ways at once, while the research on familiarity does not. The more likely culprit is the state you are in when you press play.
This post is not about whether music helps you concentrate, which is a separate question covered in the post on whether music helps you study. It is about the specific complaint that something used to work and now does not, and what the research suggests you actually change.
Why did it work in the first place?
The leading explanation is less about music than people expect.
In 2001, Thompson, Schellenberg and Husain tested the Mozart effect properly. Participants did a spatial task after listening to Mozart, which is fast and major and pleasant, after an Albinoni adagio, which is slow and sad, or after silence. Performance improved after Mozart and not after Albinoni. Then they statistically controlled for differences in arousal, mood and enjoyment, and the effect disappeared.
That is the arousal and mood account, and it has been the dominant explanation ever since. Music does not do something special to cognition. It changes how alert and how cheerful you are, and those changes show up in performance. Anything else that lifted your arousal and mood would do the same job.
The account is leading rather than settled. In 2024, Chee and colleagues published a systematic review of 31 studies that measured autonomic arousal objectively, using cardiac, electrodermal and pupil measures, alongside cognitive performance under music and other auditory stimulation. Their conclusion was that the evidence for or against the arousal and mood hypothesis is at best indirect. So the best available explanation for why music helps is also one the physiological data has not confirmed.
Hold onto the implication though, because it is the useful part: if the mechanism runs through arousal and mood, then the same music will do different things to you depending on what state you bring to it.
Do you actually habituate to a playlist?
This is where the common advice runs into evidence that does not agree with itself.
Röer, Bell, Dentale and Buchner tested habituation in 2011, repeating a distracting sound across a session to see whether its disruptive effect shrank. It did not. The size of the auditory distraction effect did not decrease with repeated exposure, and switching to a novel distractor after a long stretch of repetition produced a temporary further drop in performance, an orienting response to the new thing.
The same group then found the opposite, under a different setup. In 2012, Bell, Röer, Dentale and Buchner showed that the irrelevant sound effect is attenuated after passive listening to the distractor beforehand: 20 minutes of passive exposure abolished the effect, and 45 seconds significantly reduced it. Crucially, the attenuation appeared only when the material in the task matched the material heard during the exposure.
So habituation does happen, and it happens to the specific thing you have been listening to. Between the two papers the picture is that you do not drift into tolerance by accident during a task, but you do acclimatize to a particular sound you have already spent time with. Both of those point the same practical direction, and it is not toward a new playlist.
The honest caveat is about timescale and paradigm. These are serial-recall experiments with a distracting sound, run within a session, not studies of the same lofi station every evening for a month. Nobody has run the multi-week version, in either direction. So “you got bored of your playlist” is not a finding, and neither is its opposite. Reading these results across to your own situation is reasoning, and it is worth labeling as such.
What does the evidence say about familiar versus new music?
It favors familiar, fairly clearly.
Bidelman and Feng ran 96 young adults through a lexical semantic task in 2025 while they listened to familiar classical music, unfamiliar classical music, or non-musical environmental sound. Familiar music produced faster correct decisions and less self-reported mind wandering than unfamiliar music. The rate of mind wandering under familiar music was statistically similar to the non-music control condition, which is close to saying that familiar music cost nothing while unfamiliar music did.
That is a direct result on the question this post is about, and it runs the opposite way from “put on something new.” Unfamiliar music has structure you have not learned yet, and learning it is engagement you did not intend to spend. The station you have heard a hundred times has nothing left to predict.
This is also, incidentally, the property lofi is built around: loops that do not resolve, tracks that end without arriving anywhere, nothing to anticipate. The lofi guide goes into why the genre is constructed that way.
Then what should you change?
Look at the state, the task and the person, in that order.
Start with the state, because the arousal and mood account makes it the first suspect. The same station at nine in the evening after a walk and at one in the morning after six hours of screens is not the same input, because you are not the same input. If music stopped working, the honest question is whether anything else is working either.
Then the task. Music costs more on some work than others. The verbal, reading-heavy work is the vulnerable kind, and that is where the auditory distraction literature consistently finds the damage. If your week moved from spreadsheets to reading, the music did not change, the job did.
Then the person, which is the part no study can settle for you. The distraction literature is full of individual differences, and the practical version is that the amount of background sound you can carry is not a fixed property of the sound. Track it across a week rather than deciding it in one bad evening.
The practical sequence that follows from all of this is unglamorous. Keep the music you know. Change the session instead: when you start, how long the block is, whether there is a defined end. The deep work guide covers how to build a block that holds, and the how to enter flow state guide covers the entry conditions, which are mostly about the task being the right difficulty rather than about what is in your ears.
Then, if you still want to change the audio, change it downward rather than sideways. Quieter, duller, less. Not newer.
That is the version of this the app is built for. Pocket Lofi’s Focus station is instrumental, repetitive and deliberately unremarkable, the sort of thing you are supposed to stop hearing, and the timer means the block has a stated end before you begin. Starting the station before you start the work, rather than reaching for it after you are already stuck, is the small change that does the most.
None of this promises anything about output. The effects in the studies above are measured in reaction times and mind-wandering reports, not in work finished, and the mechanism everyone quotes is one a 2024 review found only indirect evidence for, in either direction. What the evidence does suggest is that when the music stops working, the music is rarely what broke.
Sources
- Arousal, mood, and the Mozart effect, Thompson WF, Schellenberg EG, Husain G, Psychological Science, 2001
- The effects of music and auditory stimulation on autonomic arousal, cognition and attention: A systematic review, Chee ZJ, Chang CYM, Cheong JY, Malek FHBA, Hussain S, de Vries M, Bellato A, International Journal of Psychophysiology, 2024
- Familiar Music Reduces Mind Wandering and Boosts Behavioral Performance During Lexical Semantic Processing, Bidelman GM, Feng S, Brain Sciences, 2025
- The role of habituation and attentional orienting in the disruption of short-term memory performance, Röer JP, Bell R, Dentale S, Buchner A, Memory and Cognition, 2011
- Habituation of the irrelevant sound effect: evidence for an attentional theory of short-term memory disruption, Bell R, Röer JP, Dentale S, Buchner A, Journal of Experimental Psychology: Learning, Memory, and Cognition, 2012