Dear Sirs,The etiology of chronic cluster headache (CH), manifestingas episodes of severe head pain with cranial autonomicfeatures is still unknown, but a central role of the hypo-thalamus has been discussed recently. For example, duringacute CH attacks (but not during migraine) PET revealedactivation of the ipsilateral posterior inferior hypothalamus(PIH) [9]. In addition, increased grey matter density of theipsilateral PIH has been found [7, 8]. These findings haveprompted the use of deep brain stimulation (DBS) of theipsilateral PIH in CH, which has been shown to be apromising option in cases refractory to other therapies.Currently available series suggest that the procedure isworking in about 50% of operated patients [1, 6].While suggestive, the previously described structural andfunctionalchangesofthePIHinCHleaveopenthequestionasto the neural events that precipitate and mediate spontaneousCH attacks within the hypothalamus. Spontaneous dischargeproperties of hypothalamic neurons have been characterizedby intraoperative microelectrode recordings to be dominatedby 25 Hz oscillations outside of active cluster episodes [3, 4].In addition, a recent report described changes of local fieldpotentials (LFP) recorded from a DBS electrode during anacute attack [2] with a power increaseat 20 Hz at the onset ofthe attack.Here, weextendthisobservation withdatafromanadditional case. A 52 year old man had suffered from right-sidedchronicCHsinceage41andhadreceivedprolongedbutunsuccessful courses of all standard treatments (includingoxygen, sumatriptan, verapamil, corticosteroids). The targetin the PIH was defined by CT-stereotactic surgery refined bymicroelectrode recording. The DBS electrode (Medtronicmodel 3389) was externalized for several days for test stim-ulation. Bipolar recordings (1,000 pts/s) between the twocontacts with the greatest distance as well as surface EEGactivity were obtained. During a recording session (no stim-ulation) the patient complained about an acute severe right-sided CH episode accompanied by slight ipsilateral ptosis butno conjunctival ingestion. Coinciding with the onset of theattack,theLFPspectrumshowedapronouncedpeakat20 Hz(see Figs. 1aand2). There was no such activity in the scalprecordings (Fig. 1b). Only one attack was observed. DBS ledto marked initial clinical improvement (90% decrease ofattack frequency) followed by a gradual increase of attackfrequency (to *50% of baseline) at 2 year follow-up.The present case is similar to one reported by Brittainand colleagues [2] who found a prominent peak in the LFPspectra around 16–22 Hz at the onset of an attack. The